podium

Maestro

45 of 51 criteria have a sourced finding; 41 were also cross-checked by a second agent. Every comparison that includes Maestro links here for scope, limitations, atomic answers and sources.

Humans approve the rubric, templates and prototype; they do not review every finding. An agent source cross-check is not a hands-on product test. A missing sourced finding is not evidence that a capability is absent.

Rubric
0.3 · 51 criteria
Source checks
2026-11-07 to 2026-11-13
Agent cross-checked
41 criteria
Comparisons
1

Access and agent choice

ACC-01

Client access

Agent cross-checked partial

Maestro documents cross-platform desktop and phone-browser control. Native distribution targets are macOS, Windows and Linux; phones use the built-in web interface, with no native iOS or Android target in this release. Browsers get the full live-synced app while the desktop runs; desktop OS notifications exist, phone push does not.

Scope
Surfaces: desktop, ssh-remote · Version: First-party sources checked 2026-11-07; pinned implementations retain their cited revision. · Plan: Not specified · OS: macOS, Windows, Linux · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • Only the named atomic answers are established by this source pass. Remaining atomic answers retain their documented gaps.
9 atomic answers
  • Desktop os [supported] macOS, Windows and Linux. Sources: maestro:nov07-readme, maestro:nov07-remote.
  • Mobile apps [unsupported] The complete application build targets are desktop Electron packages. Phone access uses the built-in browser client; no native phone package is shipped. Sources: maestro:gap39-clients.
  • Browser [supported] Full host UI in browser. Sources: maestro:nov07-readme, maestro:nov07-remote.
  • Mobile browser [supported] Phone layout with same agents, transcripts and settings. Sources: maestro:nov07-readme, maestro:nov07-remote.
  • Cli [supported] maestro-cli is bundled with the desktop. Sources: maestro:nov07-cli.
  • Editor extensions [unknown] Editor extensions: The November first-party documentation and repository pass did not establish client access at the rubric's required scope. No installed-product fixture was run.
  • Task actions [supported] The browser interface exposes the same agents, tabs, transcripts and Auto Run as the desktop; maestro-cli sends, dispatches, inspects, stops, resumes, skips, aborts and steers runs. Sources: maestro:nov21-remote, maestro:nov21-cli.
  • Notifications [partial] Desktop: OS notifications (task complete, long command finished, attention needed), in-app toasts and a custom command on completion. The web interface can show browser notifications while connected; no Web Push to phones was found. Sources: maestro:nov21-config-notify, maestro:nov21-features, maestro:nov21-web-notify.
  • Cross device continuity [partial] Browsers, including phones, get the full app live-synced with the desktop window and reload state after reconnecting, but only while the desktop app runs; it is the single source of truth. Sources: maestro:nov21-remote.

Sources: maestro:nov07-readme, maestro:nov07-remote, maestro:gap39-clients, maestro:nov21-remote, maestro:nov21-config-notify, maestro:nov21-features, maestro:nov21-web-notify, maestro:nov21-cli

Agent cross-checked partial

At v1.0.4 Maestro drives eleven coding CLIs as providers: Claude Code, Codex, OpenCode and Factory Droid (fully integrated); Copilot-CLI, Hermes, Pi, Qwen3 Coder and Oh My Pi (beta); and Antigravity CLI and Grok CLI, which have visible provider definitions but are not in the installation list. Gemini CLI is hidden in code and documented as planned support. Agents on different providers consult each other through mentions and Group Chat.

Scope
Surfaces: desktop, ssh-remote · Version: v1.0.4 (508ef1201846f16fdf79f3b75ce17db838aa2dbc, released 2026-11-21): src/shared/agentIds.ts, src/main/agents/definitions.ts, src/main/index.ts, docs/installation.md, docs/provider-notes.md and the collaboration docs. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • The rubric scenario is unverified. Documentation and source inspection are not a successful product test.
  • Antigravity requires its own CLI and login. Its read-only mode is not enforced by the CLI and it has no tool-disabling flag.
5 atomic answers
  • Harnesses [supported] v1.0.4 providers: Claude Code, Codex, OpenCode, Factory Droid, Copilot-CLI, Hermes, Pi, Qwen3 Coder, Oh My Pi, Antigravity CLI and Grok CLI. The installation list marks Copilot-CLI, Hermes, Pi, Qwen3 Coder and Oh My Pi as beta and omits Antigravity and Grok. Gemini CLI is hidden ("Not shipping; superseded by Antigravity") and documented as planned. Sources: maestro:ade154-agent-ids, maestro:ade154-agent-defs, maestro:ade154-installation.
  • Model providers [partial] Model choice belongs to each CLI: Claude Code uses the Anthropic account; Codex, OpenCode, Factory Droid, Copilot-CLI (multi-model via models.dev), Hermes, Pi, Qwen3 Coder, Oh My Pi and Antigravity take a model flag or setting. Local runtimes work by pointing a provider's base URL at a compatible local server. Sources: maestro:ade154-provider-notes, maestro:ade154-agent-defs, maestro:nov21-multiprov.
  • Custom adapters [unsupported] No custom adapters at v1.0.4: AGENT_IDS is a closed list extended only in code, and settings can only override the binary path of a listed agent; a search of src for "custom agent", "customAgent" and "customProvider" found no way to register a new agent type. Listed providers accept custom arguments and environment variables. Sources: maestro:ade154-agent-ids, maestro:ade154-custom-paths, maestro:ade154-provider-notes.
  • Session integration [supported] Maestro runs providers in batch mode with provider-specific JSON stream output; Claude Code can instead run each turn through its interactive TUI driven by the bundled maestro-p helper. Sources: maestro:ade154-provider-notes, maestro:ade154-installation, maestro:ade154-agent-defs.
  • Mixed harness operation [partial] Agents on different providers can consult each other through mentions and Group Chat; the Group Chat moderator must be Claude Code, OpenCode or Codex. Moving one agent to another provider drops its history (roadmap). Sources: maestro:nov21-collab, maestro:nov21-groupchat, maestro:nov21-multiprov.

Sources: maestro:src2-maestro-release-v100, maestro:nov21-collab, maestro:nov21-groupchat, maestro:nov21-multiprov, maestro:ade154-agent-ids, maestro:ade154-agent-defs, maestro:ade154-custom-paths, maestro:ade154-installation, maestro:ade154-provider-notes

Agent cross-checked partial

Provider credentials decide subscription versus API billing. No Maestro account or credits exist; accounts switch per agent through provider config variables and local models via base URLs.

Scope
Surfaces: desktop, ssh-remote · Version: First-party sources checked 2026-11-07; pinned implementations retain their cited revision. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • The remaining atomic questions and the rubric scenario are unverified. Documentation is not a successful product test.
7 atomic answers
  • Agent subscriptions [supported] Claude plans can fund compatible transports. Sources: maestro:nov07-providers.
  • Api accounts [supported] API keys/gateways/Bedrock/Vertex can fund Claude calls. Sources: maestro:nov07-providers.
  • Vendor credits [unsupported] Maestro sells no credits: the site has no pricing or plans and v1.0.4 source has no billing code; model usage is billed by the provider. Sources: maestro:nov21-home, maestro:nov21-billing-search, maestro:nov21-install.
  • Local models [partial] A provider's base-URL variable can point at Ollama or LM Studio; the endpoint must speak that CLI's API. Sources: maestro:nov21-multiprov.
  • Account switching [partial] Per agent via CLAUDE_CONFIG_DIR (Claude Code), CODEX_HOME (Codex) or XDG_DATA_HOME (OpenCode); Copilot CLI has one machine-wide login and the other beta providers have no selector. Sources: maestro:nov21-multiprov.
  • Vendor account requirement [supported] No Maestro account: install and an authenticated provider CLI suffice; only the optional leaderboard has a sign-up. Sources: maestro:nov21-install, maestro:nov21-start, maestro:nov21-achievements.
  • Provider restrictions [partial] SSH lacks Standard permission relay; TUI wrapper requires remote maestro-p. Sources: maestro:nov07-providers.

Sources: maestro:nov07-providers, maestro:nov21-multiprov, maestro:nov21-install, maestro:nov21-start, maestro:nov21-achievements, maestro:nov21-home, maestro:nov21-billing-search

No sourced public finding not-published

No publishable sourced finding is available for this criterion. This does not mean the product lacks the capability.

Agent cross-checked partial

Maestro has bundled CLI control plus a separate read-only documentation-search MCP server.

Scope
Surfaces: desktop, ssh-remote · Version: First-party sources checked 2026-11-07; pinned implementations retain their cited revision. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-07
  • The remaining atomic questions and the rubric scenario are unverified. Documentation is not a successful product test.
  • Only the named atomic facts are established by documentation. Remaining rubric scenarios and unknown answers were not demonstrated.
8 atomic answers
  • Cli [supported] maestro-cli. Sources: maestro:nov07-cli.
  • Api [unknown] Api: The November first-party documentation and repository pass did not establish programmable access at the rubric's required scope. No installed-product fixture was run.
  • Sdk [unknown] Sdk: The November first-party documentation and repository pass did not establish programmable access at the rubric's required scope. No installed-product fixture was run.
  • Mcp server [supported] SearchMaestro MCP searches documentation, not task/session operations; public read-only HTTPS endpoint. Sources: maestro:gap-docs-mcp.
  • Webhooks [unknown] Webhooks: The November first-party documentation and repository pass did not establish programmable access at the rubric's required scope. No installed-product fixture was run.
  • Operations [supported] Send/resume, list sessions, run playbooks and manage settings. Sources: maestro:nov07-cli.
  • Authentication [supported] Documentation MCP needs no authentication; provider CLI and server authentication are separate. Sources: maestro:gap-docs-mcp.
  • Stability [partial] Standardized exit codes distinguish missing app, timeout and unsupported operations. Sources: maestro:nov07-cli.

Sources: maestro:nov07-cli, maestro:gap-docs-mcp

Agent cross-checked partial

Maestro inherits MCP tools, skills and authentication from configured coding providers. Built-in and optional third-party plugins, custom slash commands and the Coworking MCP server add to provider-supplied tools.

Scope
Surfaces: desktop, ssh-remote · Version: First-party documentation/source checked 2026-11-07; exact pinned source revision is recorded in each citation. No installed-product fixture. · Plan: Desktop local/SSH; headless Cue is separately documented with unresolved release parity · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
unknown
Delivery
unknown
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • Only the named atomic facts are established by documentation. Remaining rubric scenarios and unknown answers were not demonstrated.
7 atomic answers
  • Mcp client [supported] Configured provider MCP tools work through the underlying provider; no independent Maestro tool-call test. Sources: maestro:gap-pass-through.
  • Plugins [supported] Settings > Plugins lists built-in plugins (Cue, Director's Notes, Symphony, Coworking and others) with permissions; third-party manifest plugins in three trust tiers sit behind the plugins flag, off by default. Sources: maestro:nov21-plugins, maestro:nov21-plugin-arch.
  • Custom tools [partial] Custom slash commands with template variables; tools themselves come from the provider CLI. Coworking adds a per-agent MCP server for terminal and browser access. Sources: maestro:nov21-slash, maestro:nov21-coworking.
  • Instructions import [supported] Configured provider skills are inherited. Sources: maestro:gap-pass-through.
  • Authentication [supported] Integration authentication is the provider CLI's own; Maestro passes MCP tools, skills, permissions and logins through. Sources: maestro:nov21-install.
  • Permission scope [supported] Provider-configured permissions remain provider-specific. Sources: maestro:gap-pass-through.
  • Integration owner [supported] Provider CLI owns MCP/tool configuration and authentication, with Maestro passing prompts. Sources: maestro:gap-pass-through.

Sources: maestro:gap-pass-through, maestro:nov21-install, maestro:nov21-slash, maestro:nov21-coworking, maestro:nov21-plugins, maestro:nov21-plugin-arch

Agent cross-checked partial

The desktop package uses Electron. Phones use a PWA served by the desktop app.

Scope
Surfaces: desktop · Version: First-party sources checked 2026-11-07; pinned implementations retain their cited revision. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • The remaining atomic questions and the rubric scenario are unverified. Documentation is not a successful product test.
4 atomic answers
  • Desktop implementation [supported] Electron. Sources: maestro:nov07-build.
  • Mobile implementation [supported] Phones use a web/PWA interface served by the desktop app (manifest and service worker); there is no native mobile client. Sources: maestro:nov21-pwa, maestro:nov21-remote.
  • Hybrids [unknown] Hybrids: The November first-party documentation and repository pass did not establish client implementation at the rubric's required scope. No installed-product fixture was run.
  • Implementation evidence [supported] Pinned package build scripts and dependency. Sources: maestro:nov07-build.

Sources: maestro:nov07-build, maestro:nov21-pwa, maestro:nov21-remote

Planning and context

Agent cross-checked partial

Work starts from prompts or local markdown task documents; Cue reacts to GitHub events and Symphony picks GitHub issues.

Scope
Surfaces: desktop · Version: Maestro v1.0.4 (508ef12, released 2026-11-21): pinned docs and source rechecked 2026-11-21. No installed-product fixture. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
unknown
Delivery
unknown
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • Only the cited atomic answers are established from documentation or pinned source; remaining atoms stay unknown and no hands-on test was run.
9 atomic answers
  • Local creation [supported] Auto Run documents are local markdown checkbox files, saved as playbooks. Sources: maestro:nov21-autorun.
  • Prompt intake [supported] Chat prompts start work, and the /wizard Inline Wizard turns a conversation into Auto Run documents. Sources: maestro:nov21-autorun.
  • External import [partial] Symphony picks a GitHub issue from registered open-source projects; general tracker import is not documented. Sources: maestro:nov21-symphony.
  • Chat git triggers [partial] Cue triggers on new GitHub PRs, issues and labels, file changes and timers; no chat-app trigger. Sources: maestro:nov21-cue.
  • Authoritative tracker [unknown] Authoritative tracker: The November first-party documentation and repository pass did not establish intake and external tracker relationship at the rubric's required scope. No installed-product fixture was run.
  • Stable ids [unknown] Stable ids: The November first-party documentation and repository pass did not establish intake and external tracker relationship at the rubric's required scope. No installed-product fixture was run.
  • Sync [unknown] Sync: The November first-party documentation and repository pass did not establish intake and external tracker relationship at the rubric's required scope. No installed-product fixture was run.
  • Supported edits [unknown] Supported edits: The November first-party documentation and repository pass did not establish intake and external tracker relationship at the rubric's required scope. No installed-product fixture was run.
  • External roundtrips [unknown] External roundtrips: The November first-party documentation and repository pass did not establish intake and external tracker relationship at the rubric's required scope. No installed-product fixture was run.

Sources: maestro:nov21-cue, maestro:nov21-symphony, maestro:nov21-autorun

Agent cross-checked partial

Auto Run uses editable Markdown plans with on-disk checkboxes and resettable run copies.

Scope
Surfaces: desktop, ssh-remote · Version: First-party sources checked 2026-11-07; pinned implementations retain their cited revision. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • The remaining atomic questions and the rubric scenario are unverified. Documentation is not a successful product test.
  • Only the named atomic facts are established by documentation. Remaining rubric scenarios and unknown answers were not demonstrated.
7 atomic answers
  • Plans [supported] Markdown task documents and reusable playbooks. Sources: maestro:nov07-autorun.
  • Acceptance criteria [unknown] Acceptance criteria: The November first-party documentation and repository pass did not establish durable task plans at the rubric's required scope. No installed-product fixture was run.
  • Editors [partial] Editor disabled during active execution, enabled while paused for an error or human gate. Sources: maestro:nov07-autorun.
  • Revisions [unknown] Revisions: The November first-party documentation and repository pass did not establish durable task plans at the rubric's required scope. No installed-product fixture was run.
  • Amendment [supported] Document editing/checkbox changes reopen while paused on errors or human gates. Sources: maestro:gap-auto.
  • Restart persistence [supported] Markdown tasks are stored on disk; Reset on Completion keeps working copies in runs/ as task records. Sources: maestro:gap-auto.

Sources: maestro:nov07-autorun, maestro:gap-auto, maestro:nov21-history

Agent cross-checked partial

Playbooks order documents and worktree agents nest under a parent; no workflow is mandatory.

Scope
Surfaces: desktop · Version: Maestro v1.0.4 (508ef12, released 2026-11-21): pinned docs and source rechecked 2026-11-21. No installed-product fixture. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
unknown
Delivery
unknown
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • Only the cited atomic answers are established from documentation or pinned source; remaining atoms stay unknown and no hands-on test was run.
9 atomic answers
  • Epics [unknown] Epics: The November first-party documentation and repository pass did not establish hierarchy, dependencies and evolving scope at the rubric's required scope. No installed-product fixture was run.
  • Hierarchy [partial] A playbook is an ordered set of documents; worktree agents nest under their parent agent. No epics or dependency enforcement. Sources: maestro:nov21-autorun, maestro:nov21-worktrees.
  • Blockers [unknown] Blockers: The November first-party documentation and repository pass did not establish hierarchy, dependencies and evolving scope at the rubric's required scope. No installed-product fixture was run.
  • Readiness [unknown] Readiness: The November first-party documentation and repository pass did not establish hierarchy, dependencies and evolving scope at the rubric's required scope. No installed-product fixture was run.
  • Enforcement [unknown] Enforcement: The November first-party documentation and repository pass did not establish hierarchy, dependencies and evolving scope at the rubric's required scope. No installed-product fixture was run.
  • State propagation [unknown] State propagation: The November first-party documentation and repository pass did not establish hierarchy, dependencies and evolving scope at the rubric's required scope. No installed-product fixture was run.
  • Task to epic [unknown] Task to epic: The November first-party documentation and repository pass did not establish hierarchy, dependencies and evolving scope at the rubric's required scope. No installed-product fixture was run.
  • Identity retention [unknown] Identity retention: The November first-party documentation and repository pass did not establish hierarchy, dependencies and evolving scope at the rubric's required scope. No installed-product fixture was run.
  • Workflow constraints [supported] No mandatory workflow: chat needs none, and Auto Run is Spec-Driven or Goal-Driven per run. Sources: maestro:nov21-autorun.

Sources: maestro:nov21-autorun, maestro:nov21-worktrees

Agent cross-checked partial

Maestro exposes Claude Code project memory for reading, editing and deletion, and passes provider-configured skills through to the underlying CLI.

Scope
Surfaces: desktop, ssh-remote · Version: First-party sources checked 2026-11-07; pinned implementations retain their cited revision. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Claude Code
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • The remaining atomic questions and the rubric scenario are unverified. Documentation is not a successful product test.
  • Only the named atomic facts are established by documentation. Remaining rubric scenarios and unknown answers were not demonstrated.
10 atomic answers
  • Instructions [partial] README says configured provider skills, MCP tools, permissions and authentication carry through; instruction files are not named. Sources: maestro:gap-pass-through.
  • Skills [supported] Provider-configured skills are passed through rather than reimplemented. Sources: maestro:gap-pass-through.
  • Project docs [unknown] Project docs: The November first-party documentation and repository pass did not establish shared context and memory at the rubric's required scope. No installed-product fixture was run.
  • Memory [supported] Claude Code project Markdown memories. Sources: maestro:nov07-memories.
  • Scope [supported] Claude Code only; other provider panels hidden. Sources: maestro:nov07-memories.
  • Provenance [unknown] Provenance: The November first-party documentation and repository pass did not establish shared context and memory at the rubric's required scope. No installed-product fixture was run.
  • Correction [supported] Preview/Edit and Save. Sources: maestro:nov07-memories.
  • Deletion [supported] Delete stale entries. Sources: maestro:nov07-memories.
  • Freshness [unknown] Freshness: The November first-party documentation and repository pass did not establish shared context and memory at the rubric's required scope. No installed-product fixture was run.
  • Takeover continuity [partial] Switching an agent to another provider drops its history and memories; Claude Code accounts can share sessions through symlinks. Sources: maestro:nov21-multiprov.

Sources: maestro:nov07-memories, maestro:gap-pass-through, maestro:nov21-multiprov

Agent cross-checked partial

File browsing and @ mentions plus image inputs; code search comes from the provider CLI.

Scope
Surfaces: desktop · Version: Maestro v1.0.4 (508ef12, released 2026-11-21): pinned docs and source rechecked 2026-11-21. No installed-product fixture. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
unknown
Delivery
unknown
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • Only the cited atomic answers are established from documentation or pinned source; remaining atoms stay unknown and no hands-on test was run.
10 atomic answers
  • Indexing [unknown] Indexing: The November first-party documentation and repository pass did not establish architecture and api awareness at the rubric's required scope. No installed-product fixture was run.
  • External knowledge [unknown] External knowledge: The November first-party documentation and repository pass did not establish architecture and api awareness at the rubric's required scope. No installed-product fixture was run.
  • Multimodal inputs [supported] Pasted and annotated images are sent with prompts. Sources: maestro:nov21-features.
  • Freshness [unknown] Freshness: The November first-party documentation and repository pass did not establish architecture and api awareness at the rubric's required scope. No installed-product fixture was run.
  • Access scope [unknown] Access scope: The November first-party documentation and repository pass did not establish architecture and api awareness at the rubric's required scope. No installed-product fixture was run.
  • Api retrieval [unknown] Api retrieval: The November first-party documentation and repository pass did not establish architecture and api awareness at the rubric's required scope. No installed-product fixture was run.
  • Citations [unknown] Citations: The November first-party documentation and repository pass did not establish architecture and api awareness at the rubric's required scope. No installed-product fixture was run.
  • Conflict detection [unknown] Conflict detection: The November first-party documentation and repository pass did not establish architecture and api awareness at the rubric's required scope. No installed-product fixture was run.
  • Review enforcement [unknown] Review enforcement: The November first-party documentation and repository pass did not establish architecture and api awareness at the rubric's required scope. No installed-product fixture was run.

Sources: maestro:nov21-features

Agent cross-checked partial

Markdown tasks can be saved and selected before explicit Auto Run.

Scope
Surfaces: desktop, ssh-remote · Version: First-party sources checked 2026-11-07; pinned implementations retain their cited revision. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • The remaining atomic questions and the rubric scenario are unverified. Documentation is not a successful product test.
17 atomic answers
  • Create [partial] Users create Auto Run documents and playbooks, not issues. Sources: maestro:nov21-autorun.
  • Edit [unknown] Edit: The November first-party documentation and repository pass did not establish native issue management and idea backlog at the rubric's required scope. No installed-product fixture was run.
  • Discuss [unknown] Discuss: The November first-party documentation and repository pass did not establish native issue management and idea backlog at the rubric's required scope. No installed-product fixture was run.
  • Prioritize [unknown] Prioritize: The November first-party documentation and repository pass did not establish native issue management and idea backlog at the rubric's required scope. No installed-product fixture was run.
  • Label [unknown] Label: The November first-party documentation and repository pass did not establish native issue management and idea backlog at the rubric's required scope. No installed-product fixture was run.
  • Assign [unknown] Assign: The November first-party documentation and repository pass did not establish native issue management and idea backlog at the rubric's required scope. No installed-product fixture was run.
  • Filter [unknown] Filter: The November first-party documentation and repository pass did not establish native issue management and idea backlog at the rubric's required scope. No installed-product fixture was run.
  • Archive [unknown] Archive: The November first-party documentation and repository pass did not establish native issue management and idea backlog at the rubric's required scope. No installed-product fixture was run.
  • Reopen [unknown] Reopen: The November first-party documentation and repository pass did not establish native issue management and idea backlog at the rubric's required scope. No installed-product fixture was run.
  • Project epic views [unknown] Project epic views: The November first-party documentation and repository pass did not establish native issue management and idea backlog at the rubric's required scope. No installed-product fixture was run.
  • Idea parking [partial] Documents can exist before running, rather than a verified multi-user issue backlog. Sources: maestro:nov07-autorun.
  • Explicit dispatch [supported] Run dialog and Go start selected documents. Sources: maestro:nov07-autorun.
  • Task session separation [unknown] Task session separation: The November first-party documentation and repository pass did not establish native issue management and idea backlog at the rubric's required scope. No installed-product fixture was run.
  • Metadata [unknown] Metadata: The November first-party documentation and repository pass did not establish native issue management and idea backlog at the rubric's required scope. No installed-product fixture was run.
  • Durable states [partial] Checkboxes update on disk; issue lifecycle semantics unverified. Sources: maestro:nov07-autorun.
  • External tracker requirement [unknown] External tracker requirement: The November first-party documentation and repository pass did not establish native issue management and idea backlog at the rubric's required scope. No installed-product fixture was run.

Sources: maestro:nov07-autorun, maestro:nov21-autorun

Agent cross-checked partial

Agents create and run Auto Run documents via the CLI; no issue records exist.

Scope
Surfaces: desktop · Version: Maestro v1.0.4 (508ef12, released 2026-11-21): pinned docs and source rechecked 2026-11-21. No installed-product fixture. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
unknown
Delivery
unknown
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • Only the cited atomic answers are established from documentation or pinned source; remaining atoms stay unknown and no hands-on test was run.
14 atomic answers
  • Shared ids [unknown] Shared ids: The November first-party documentation and repository pass did not establish agent-native issue operations at the rubric's required scope. No installed-product fixture was run.
  • Read [unknown] Read: The November first-party documentation and repository pass did not establish agent-native issue operations at the rubric's required scope. No installed-product fixture was run.
  • Create [partial] Agents can write Auto Run documents and run them through maestro-cli; there are no issue records. Sources: maestro:nov21-cli, maestro:nov21-autorun.
  • Update [unknown] Update: The November first-party documentation and repository pass did not establish agent-native issue operations at the rubric's required scope. No installed-product fixture was run.
  • Claim [unknown] Claim: The November first-party documentation and repository pass did not establish agent-native issue operations at the rubric's required scope. No installed-product fixture was run.
  • Assign [unknown] Assign: The November first-party documentation and repository pass did not establish agent-native issue operations at the rubric's required scope. No installed-product fixture was run.
  • State changes [unknown] State changes: The November first-party documentation and repository pass did not establish agent-native issue operations at the rubric's required scope. No installed-product fixture was run.
  • Dependencies [unknown] Dependencies: The November first-party documentation and repository pass did not establish agent-native issue operations at the rubric's required scope. No installed-product fixture was run.
  • Children [unknown] Children: The November first-party documentation and repository pass did not establish agent-native issue operations at the rubric's required scope. No installed-product fixture was run.
  • Results [unknown] Results: The November first-party documentation and repository pass did not establish agent-native issue operations at the rubric's required scope. No installed-product fixture was run.
  • Human visibility [unknown] Human visibility: The November first-party documentation and repository pass did not establish agent-native issue operations at the rubric's required scope. No installed-product fixture was run.
  • Attribution [unknown] Attribution: The November first-party documentation and repository pass did not establish agent-native issue operations at the rubric's required scope. No installed-product fixture was run.
  • Permissions [unknown] Permissions: The November first-party documentation and repository pass did not establish agent-native issue operations at the rubric's required scope. No installed-product fixture was run.
  • Conflicts [unknown] Conflicts: The November first-party documentation and repository pass did not establish agent-native issue operations at the rubric's required scope. No installed-product fixture was run.

Sources: maestro:nov21-cli, maestro:nov21-autorun

Agent cross-checked partial

Spec documents persist as markdown files without requirement links.

Scope
Surfaces: desktop · Version: Maestro v1.0.4 (508ef12, released 2026-11-21): pinned docs and source rechecked 2026-11-21. No installed-product fixture. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
unknown
Delivery
unknown
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • Only the cited atomic answers are established from documentation or pinned source; remaining atoms stay unknown and no hands-on test was run.
9 atomic answers
  • Persistent requirements [partial] Spec documents persist as markdown files (with bundled Spec-Kit and OpenSpec commands); no requirement links or drift checks. Sources: maestro:nov21-autorun, maestro:nov21-slash.
  • Human edits [unknown] Human edits: The November first-party documentation and repository pass did not establish living product specifications at the rubric's required scope. No installed-product fixture was run.
  • Agent edits [unknown] Agent edits: The November first-party documentation and repository pass did not establish living product specifications at the rubric's required scope. No installed-product fixture was run.
  • Revisions [unknown] Revisions: The November first-party documentation and repository pass did not establish living product specifications at the rubric's required scope. No installed-product fixture was run.
  • Approvals [unknown] Approvals: The November first-party documentation and repository pass did not establish living product specifications at the rubric's required scope. No installed-product fixture was run.
  • Later retrieval [unknown] Later retrieval: The November first-party documentation and repository pass did not establish living product specifications at the rubric's required scope. No installed-product fixture was run.
  • Drift detection [unknown] Drift detection: The November first-party documentation and repository pass did not establish living product specifications at the rubric's required scope. No installed-product fixture was run.
  • Review integration [unknown] Review integration: The November first-party documentation and repository pass did not establish living product specifications at the rubric's required scope. No installed-product fixture was run.

Sources: maestro:nov21-autorun, maestro:nov21-slash

Coordination and automation

COORD-01

Parallel work

Agent cross-checked partial

Maestro documents multiple agent processes and worktree agents. Parallel Auto Runs use separate agents or worktrees; Cue caps concurrent runs per agent.

Scope
Surfaces: desktop, ssh-remote · Version: First-party sources checked 2026-11-07; pinned implementations retain their cited revision. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • The remaining atomic questions and the rubric scenario are unverified. Documentation is not a successful product test.
5 atomic answers
  • Concurrent sessions [supported] Multiple active agents. Sources: maestro:nov07-readme.
  • Concurrent tasks [supported] Auto Runs run in parallel across agents, and in one repository through worktree sub-agents with separate directories and branches. Sources: maestro:nov21-autorun, maestro:nov21-worktrees.
  • Quotas [partial] No product-wide cap is documented; Cue allows 1-10 simultaneous runs per agent (default 1) and queues up to queue_size events. Sources: maestro:nov21-cue-config.
  • Observed concurrency [unknown] Observed concurrency: The November first-party documentation and repository pass did not establish parallel work at the rubric's required scope. No installed-product fixture was run.
  • Status visibility [partial] Agent status dots show Ready, Thinking and Error; status correctness was not tested. Sources: maestro:nov07-config.

Sources: maestro:nov07-readme, maestro:nov07-config, maestro:nov21-autorun, maestro:nov21-worktrees, maestro:nov21-cue-config

Agent cross-checked partial

Group Chat queues busy agents by default, with an explicit unsafe concurrent-edit alternative. Users and Cue subscriptions assign agents; the Group Chat moderator routes work.

Scope
Surfaces: desktop, ssh-remote · Version: First-party documentation/source checked 2026-11-07; exact pinned source revision is recorded in each citation. No installed-product fixture. · Plan: Desktop local/SSH; headless Cue is separately documented with unresolved release parity · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
unknown
Delivery
unknown
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • Only the named atomic facts are established by documentation. Remaining rubric scenarios and unknown answers were not demonstrated.
7 atomic answers
  • Human assignment [unknown] Human assignment: The November first-party documentation and repository pass did not establish ownership and dispatch at the rubric's required scope. No installed-product fixture was run.
  • Agent assignment [supported] The user picks the agent for each run; Cue subscriptions route events to configured target agents, with fan-out to several. Sources: maestro:nov21-autorun, maestro:nov21-cue.
  • Claims [unknown] Claims: The November first-party documentation and repository pass did not establish ownership and dispatch at the rubric's required scope. No installed-product fixture was run.
  • Queues [supported] Only work with agents that are free waits for busy agents up to 15 minutes, then moves on. Sources: maestro:gap-group.
  • Routing [partial] Group Chat's moderator routes the question to participants by @mention; there is no rule-based routing beyond Cue targets. Sources: maestro:nov21-groupchat, maestro:nov21-collab.
  • Duplicate prevention [supported] Default busy-agent check avoids concurrent calls to an agent already in own chat, Auto Run or CLI; disabling permits two processes without locks. Sources: maestro:gap-group.
  • Ownership enforcement [unknown] Ownership enforcement: The November first-party documentation and repository pass did not establish ownership and dispatch at the rubric's required scope. No installed-product fixture was run.

Sources: maestro:gap-group, maestro:nov21-autorun, maestro:nov21-cue, maestro:nov21-groupchat, maestro:nov21-collab

Agent cross-checked partial

Group Chat has a moderator, follow-up rounds and named local/remote participants.

Scope
Surfaces: desktop, ssh-remote · Version: First-party sources checked 2026-11-07; pinned implementations retain their cited revision. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-07
  • The remaining atomic questions and the rubric scenario are unverified. Documentation is not a successful product test.
  • Only the named atomic facts are established by documentation. Remaining rubric scenarios and unknown answers were not demonstrated.
12 atomic answers
  • Roles [supported] Moderator and participants. Sources: maestro:nov07-group.
  • Model harness effort [supported] Group chat Edit changes moderator and its CLI args, path and environment. Sources: maestro:gap-group.
  • Messages [supported] Moderator routes @mentions. Sources: maestro:nov07-group.
  • Results return [supported] Moderator synthesizes participant answers. Sources: maestro:nov07-group.
  • Cross harness [supported] Moderator can be Claude Code, OpenCode or Codex with participants as Maestro sessions; no mixed-harness fixture. Sources: maestro:gap-group.
  • Orchestration owner [supported] Maestro Group Chat. Sources: maestro:nov07-group.
  • Activation [unknown] Activation: The November first-party documentation and repository pass did not establish inspectable delegation and communication at the rubric's required scope. No installed-product fixture was run.
  • Authentication [unknown] Authentication: The November first-party documentation and repository pass did not establish inspectable delegation and communication at the rubric's required scope. No installed-product fixture was run.
  • Live progress [unknown] Live progress: The November first-party documentation and repository pass did not establish inspectable delegation and communication at the rubric's required scope. No installed-product fixture was run.
  • Retained results [supported] Group chat conversation history persists until deleted. Sources: maestro:gap-group.
  • Followup [supported] Moderator can request another round. Sources: maestro:nov07-group.
  • Invocation [supported] Create group, select moderator and @mention agents. Sources: maestro:nov07-group.

Sources: maestro:nov07-group, maestro:gap-group

Agent cross-checked partial

Released desktop Cue sends prompts on file changes, timers, scheduled times, agent completions and GitHub events. Online headless-server documentation describes an additional deployment whose parity with this pinned release remains unestablished. Released Cue needs the desktop app running; goal and Cue runs have stop conditions, and provider failures are retried.

Scope
Surfaces: desktop · Version: v1.0.4 (508ef1201846f16fdf79f3b75ce17db838aa2dbc, released 2026-11-21) (released desktop Cue; headless parity unresolved) · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • The release source establishes desktop Cue. Headless-server parity remains unestablished and is excluded from the current availability claim; the published server guide exists only on the rc branch at v1.0.4.
8 atomic answers
  • Schedules [supported] Desktop Cue time.heartbeat and time.scheduled provide interval and scheduled triggers. Sources: maestro:ade154-cue-triggers.
  • Events [supported] Desktop Cue subscriptions react to file changes, agent completions, GitHub activity and webhooks. Sources: maestro:ade154-cue-overview, maestro:ade154-cue-triggers.
  • Recurring tasks [supported] Cue time.heartbeat subscriptions repeat at a set interval and time.scheduled fires at set times and weekdays, in the released desktop app (v1.0.4). Not tested. Sources: maestro:ade154-cue-triggers.
  • Completion conditions [partial] Goal runs stop on completion, deadlock, max iterations or stall; Cue runs time out after timeout_minutes (default 30) with break or continue. Sources: maestro:nov21-autorun, maestro:nov21-cue-config.
  • Scheduler location [supported] The established scheduler is the released desktop Cue engine. Sources: maestro:ade154-cue-overview.
  • Overlaps [supported] Per-agent max_concurrent (default 1) serialises Cue runs; extra events queue up to queue_size (default 512) and overflow is dropped with a notice. Not tested. Sources: maestro:ade154-cue-concurrency, maestro:nov21-cue-config.
  • Retries [partial] Agent Resilience resends Cue, dispatch and web turns on provider overload or quota errors and resumes Auto Run batches; CLI batches do not retry and pending retries stop when Maestro quits. Sources: maestro:nov21-resilience.
  • Client independence [unsupported] Released v1.0.4 Cue runs inside the desktop app: a one-shot fire missed while Maestro is closed runs on next start within a 360-minute default grace window. The published "Running Cue on a Server" page exists only on the rc branch, not in this release. CLI headless runs need the user's own cron. Sources: maestro:nov21-cue-events, maestro:nov21-cue-server-rc, maestro:nov21-cli.

Sources: maestro:nov07-cue-server, maestro:gap-server, maestro:nov21-cue-events, maestro:nov21-cue-server-rc, maestro:nov21-cli, maestro:nov21-autorun, maestro:nov21-cue-config, maestro:nov21-resilience, maestro:ade154-cue-overview, maestro:ade154-cue-triggers, maestro:ade154-cue-concurrency

No sourced public finding not-published

No publishable sourced finding is available for this criterion. This does not mean the product lacks the capability.

Agent cross-checked partial

Nested worktree agents expose parent relationships; the fleet fixture remains untested. Director's Notes shows all agents' history in one timeline.

Scope
Surfaces: desktop, ssh-remote · Version: First-party sources checked 2026-11-07; pinned implementations retain their cited revision. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • The remaining atomic questions and the rubric scenario are unverified. Documentation is not a successful product test.
8 atomic answers
  • Fleet fixture [unknown] Fleet fixture: The November first-party documentation and repository pass did not establish fleet and epic management ui at the rubric's required scope. No installed-product fixture was run.
  • Parent child [supported] Worktree agents nested below parent. Sources: maestro:nov07-git.
  • Dependencies [unknown] Dependencies: The November first-party documentation and repository pass did not establish fleet and epic management ui at the rubric's required scope. No installed-product fixture was run.
  • Attention [unknown] Attention: The November first-party documentation and repository pass did not establish fleet and epic management ui at the rubric's required scope. No installed-product fixture was run.
  • History [supported] Director's Notes aggregates every agent's History into one searchable, filterable timeline with synopses. Sources: maestro:nov21-director, maestro:nov21-history.
  • Workload [supported] Independent directory, conversation and state per worktree. Sources: maestro:nov07-git.
  • Navigation errors [unknown] Navigation errors: The November first-party documentation and repository pass did not establish fleet and epic management ui at the rubric's required scope. No installed-product fixture was run.
  • Fixture mode [unknown] Fixture mode: The November first-party documentation and repository pass did not establish fleet and epic management ui at the rubric's required scope. No installed-product fixture was run.

Sources: maestro:nov07-git, maestro:nov21-director, maestro:nov21-history

Agent cross-checked partial

Auto Run supports presets, model options and queued steering notes between tasks. Group Chat roles are visible and History keeps agent-to-agent consults.

Scope
Surfaces: desktop, ssh-remote · Version: First-party sources checked 2026-11-07; pinned implementations retain their cited revision. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • The remaining atomic questions and the rubric scenario are unverified. Documentation is not a successful product test.
  • Only the named atomic facts are established by documentation. Remaining rubric scenarios and unknown answers were not demonstrated.
10 atomic answers
  • Quick task [partial] Interactive sessions are separate from Auto Run. Growing epic fixture not run. Sources: maestro:nov07-autorun.
  • Growing epic [unknown] Growing epic: The November first-party documentation and repository pass did not establish workflow flexibility and orchestration visibility at the rubric's required scope. No installed-product fixture was run.
  • Presets [supported] Spec-Driven and Goal-Driven run modes. Sources: maestro:nov07-autorun.
  • Natural language [unknown] Natural language: The November first-party documentation and repository pass did not establish workflow flexibility and orchestration visibility at the rubric's required scope. No installed-product fixture was run.
  • Agent configuration [supported] Per-run model/effort override; agent defaults retained. Sources: maestro:nov07-autorun.
  • In run changes [supported] Steer Auto Run notes are delivered at the start of the next task, with pending/delivered state visible. Sources: maestro:gap-auto.
  • Live roles [partial] Group Chat shows an appointed moderator and participants added on mention; other runs have no role view. Sources: maestro:nov21-groupchat.
  • Retained relationships [partial] History agent entries name the calling agent and question and link to the answering conversation. Sources: maestro:nov21-history.
  • Discoverability [unknown] Discoverability: The November first-party documentation and repository pass did not establish workflow flexibility and orchestration visibility at the rubric's required scope. No installed-product fixture was run.
  • Auth friction [unknown] Auth friction: The November first-party documentation and repository pass did not establish workflow flexibility and orchestration visibility at the rubric's required scope. No installed-product fixture was run.

Sources: maestro:nov07-autorun, maestro:gap-auto, maestro:nov21-groupchat, maestro:nov21-history

Execution and reliability

Sourced partial

Desktop/SSH execution and headless Cue server pipelines are user-operated. The scoped launcher has no vendor-managed runtime.

Scope
Surfaces: desktop, ssh-remote, Maestro Cue server · Version: First-party sources checked 2026-11-07; pinned implementations retain their cited revision. · Plan: Desktop and SSH execution; online headless Cue guide has unresolved parity with the pinned release source. · OS: Not specified · Execution modes: local, user-owned-ssh · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
Not separately cross-checked
  • Only the named atomic answers are established by this source pass. Remaining atomic answers retain their documented gaps.
12 atomic answers
  • Local [supported] Local desktop host. Sources: maestro:nov07-ssh, maestro:nov07-remote.
  • Byo remote [supported] SSH host executes agents. Sources: maestro:nov07-ssh, maestro:nov07-remote.
  • Self hosted [supported] Exported Cue pipeline/agents run under Linux systemd or container with no desktop window/logged-in user. Sources: maestro:gap-server.
  • Managed cloud [unsupported] Execution uses local host processes or user-supplied SSH hosts; the inspected launch paths do not provision a Maestro cloud runtime. Sources: maestro:gap39-spawn.
  • Hybrid [unknown] Hybrid: The November first-party documentation and repository pass did not establish laptop-independent execution at the rubric's required scope. No installed-product fixture was run.
  • Provisioning [partial] User configures host and agent CLIs. Sources: maestro:nov07-ssh, maestro:nov07-remote.
  • Payer [supported] User supplies Linux/container host, credentials and provider billing. Sources: maestro:gap-server.
  • Resources [unknown] Resources: The November first-party documentation and repository pass did not establish laptop-independent execution at the rubric's required scope. No installed-product fixture was run.
  • Region [unknown] Region: The November first-party documentation and repository pass did not establish laptop-independent execution at the rubric's required scope. No installed-product fixture was run.
  • Expiry [unknown] Expiry: The November first-party documentation and repository pass did not establish laptop-independent execution at the rubric's required scope. No installed-product fixture was run.
  • Client disconnect [supported] Headless server engine owns execution independently of a desktop client. Sources: maestro:gap-server.
  • Other device continuation [partial] Browser controls the running desktop host. No disconnect fixture was run. Sources: maestro:nov07-ssh, maestro:nov07-remote.

Sources: maestro:nov07-ssh, maestro:nov07-remote, maestro:gap-server, maestro:gap39-spawn

Agent cross-checked partial

Worktree agents use separate directories and branches. The scoped launcher starts local or SSH processes without provisioning per-agent containers or VMs.

Scope
Surfaces: desktop, ssh-remote · Version: First-party sources checked 2026-11-07; pinned implementations retain their cited revision. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • Only the named atomic answers are established by this source pass. Remaining atomic answers retain their documented gaps.
12 atomic answers
  • Checkout [unknown] Checkout: The November first-party documentation and repository pass did not establish workspace and process isolation at the rubric's required scope. No installed-product fixture was run.
  • Branch [supported] Worktree branch. Sources: maestro:nov07-git.
  • Worktree [supported] Separate Git worktree directory. Sources: maestro:nov07-git.
  • Container [unsupported] The scoped process spawners and runners launch host processes or SSH commands, without creating a per-agent container. Sources: maestro:gap39-spawn.
  • Vm [unsupported] The scoped process spawners and runners launch host processes or SSH commands, without creating a per-agent vm. Sources: maestro:gap39-spawn.
  • Os sandbox [unknown] Os sandbox: The November first-party documentation and repository pass did not establish workspace and process isolation at the rubric's required scope. No installed-product fixture was run.
  • Filesystem [partial] Code separation, not evidence of hostile-code containment. Sources: maestro:nov07-git.
  • Process [unknown] Process: The November first-party documentation and repository pass did not establish workspace and process isolation at the rubric's required scope. No installed-product fixture was run.
  • Network [unknown] Network: The November first-party documentation and repository pass did not establish workspace and process isolation at the rubric's required scope. No installed-product fixture was run.
  • Secrets [unknown] Secrets: The November first-party documentation and repository pass did not establish workspace and process isolation at the rubric's required scope. No installed-product fixture was run.
  • Code separation [supported] Worktree agents each have their own directory and branch. Sources: maestro:nov21-worktrees.
  • Security boundary [unknown] Security boundary: The November first-party documentation and repository pass did not establish workspace and process isolation at the rubric's required scope. No installed-product fixture was run.

Sources: maestro:nov07-git, maestro:gap39-spawn, maestro:nov21-worktrees

Agent cross-checked partial

Agent Resilience pending retries do not survive quitting the app; Cue engine restarts mark in-flight runs failed, and Group Chat queues come back paused.

Scope
Surfaces: desktop, ssh-remote · Version: First-party sources checked 2026-11-07; pinned implementations retain their cited revision. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • The remaining atomic questions and the rubric scenario are unverified. Documentation is not a successful product test.
  • Only the named atomic facts are established by documentation. Remaining rubric scenarios and unknown answers were not demonstrated.
  • service_restart and state_reconstruction come from the headless Cue server guide, which is absent from the pinned release source; released parity is unresolved.
12 atomic answers
  • Client disconnect [partial] Pending retries do not survive quitting Maestro. Sources: maestro:nov07-resilience.
  • Worker crash [unknown] Worker crash: The November first-party documentation and repository pass did not establish recovery, expiry and continuity at the rubric's required scope. No installed-product fixture was run.
  • Service restart [supported] Cue run in flight when engine stops is listed failed after engine restart; output is in logs. Sources: maestro:gap-server.
  • Expiry [unknown] Expiry: The November first-party documentation and repository pass did not establish recovery, expiry and continuity at the rubric's required scope. No installed-product fixture was run.
  • Transcript reload [supported] Sessions and History persist, and session discovery imports and resumes existing provider sessions. Sources: maestro:nov21-features, maestro:nov21-history.
  • Process reconnect [unknown] Process reconnect: The November first-party documentation and repository pass did not establish recovery, expiry and continuity at the rubric's required scope. No installed-product fixture was run.
  • Task resume [partial] Retry sends original prompt/settings; no crash recovery fixture. Sources: maestro:nov07-resilience.
  • State reconstruction [supported] Cue run/activity records persist in cue.db; stopped engine does not reconstruct a live agent turn. Sources: maestro:gap-server.
  • Context loss [unknown] Context loss: The November first-party documentation and repository pass did not establish recovery, expiry and continuity at the rubric's required scope. No installed-product fixture was run.
  • Duplicate actions [unknown] Duplicate actions: The November first-party documentation and repository pass did not establish recovery, expiry and continuity at the rubric's required scope. No installed-product fixture was run.
  • Manual repair [unknown] Manual repair: The November first-party documentation and repository pass did not establish recovery, expiry and continuity at the rubric's required scope. No installed-product fixture was run.
  • Multi day continuity [unknown] Multi day continuity: The November first-party documentation and repository pass did not establish recovery, expiry and continuity at the rubric's required scope. No installed-product fixture was run.

Sources: maestro:nov07-resilience, maestro:gap-server, maestro:gap-group, maestro:nov21-features, maestro:nov21-history

Agent cross-checked partial

Worktree setup scripts prepare missing files before agents start.

Scope
Surfaces: desktop, ssh-remote · Version: First-party sources checked 2026-11-07; pinned implementations retain their cited revision. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • The remaining atomic questions and the rubric scenario are unverified. Documentation is not a successful product test.
10 atomic answers
  • Dependencies [supported] User setup script can bootstrap ignored files/dependencies. Sources: maestro:nov07-git.
  • Caches [unknown] The checked first-party sources did not establish caches at this criterion's scope; no installed fixture was run.
  • Snapshots [unknown] Snapshots: The November first-party documentation and repository pass did not establish development environment reliability at the rubric's required scope. No installed-product fixture was run.
  • Services [unknown] Services: The November first-party documentation and repository pass did not establish development environment reliability at the rubric's required scope. No installed-product fixture was run.
  • Ports [unknown] Ports: The November first-party documentation and repository pass did not establish development environment reliability at the rubric's required scope. No installed-product fixture was run.
  • Secrets [partial] Global environment variables (e.g. API keys) apply to all agent processes and terminals; no secret store or per-workspace scoping is documented. Sources: maestro:nov21-features.
  • Duplicate installs [unknown] Duplicate installs: The November first-party documentation and repository pass did not establish development environment reliability at the rubric's required scope. No installed-product fixture was run.
  • Collisions [unknown] Collisions: The November first-party documentation and repository pass did not establish development environment reliability at the rubric's required scope. No installed-product fixture was run.
  • Remote setup [supported] Same setup runs on SSH host. Sources: maestro:nov07-git.
  • Automatic repair [unknown] Automatic repair: The November first-party documentation and repository pass did not establish development environment reliability at the rubric's required scope. No installed-product fixture was run.

Sources: maestro:nov07-git, maestro:nov21-features

Sourced partial

Maestro documents provider retries, per-run auto-resume caps and stalled-goal/document stops.

Scope
Surfaces: desktop, ssh-remote · Version: First-party sources checked 2026-11-07; pinned implementations retain their cited revision. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
Not separately cross-checked
  • The remaining atomic questions and the rubric scenario are unverified. Documentation is not a successful product test.
  • Only the named atomic facts are established by documentation. Remaining rubric scenarios and unknown answers were not demonstrated.
8 atomic answers
  • Time limits [unknown] Time limits: The November first-party documentation and repository pass did not establish limits and failure handling at the rubric's required scope. No installed-product fixture was run.
  • Turn limits [supported] Goal mode finite iteration limit can stop before completion. Sources: maestro:gap-auto.
  • Token limits [unknown] Token limits: The November first-party documentation and repository pass did not establish limits and failure handling at the rubric's required scope. No installed-product fixture was run.
  • Spend limits [unknown] Spend limits: The November first-party documentation and repository pass did not establish limits and failure handling at the rubric's required scope. No installed-product fixture was run.
  • Timeouts [supported] Watchdog/time-budget failure immediately trips the desktop document stall threshold; exact time cap unspecified. Sources: maestro:gap-auto.
  • Retries [supported] Backoff for overload; quota probes until recovery. Sources: maestro:nov07-resilience.
  • Provider failures [supported] Credential, permission, session-not-found and agent-crash errors are not retried. Sources: maestro:nov07-resilience.
  • Enforcement [partial] Visible outage card has Try now and Stop; no bounded failure fixture. Sources: maestro:nov07-resilience.

Sources: maestro:nov07-resilience, maestro:gap-auto

Agent cross-checked partial

The Usage Dashboard shows per-agent, group, Auto Run, Cue, token and time views with JSON/CSV export that includes tokens and cost; History, Director's Notes and the Cue activity log record agent activity.

Scope
Surfaces: desktop, ssh-remote · Version: v1.0.4 (508ef1201846f16fdf79f3b75ce17db838aa2dbc, released 2026-11-21) · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • The remaining atomic questions and the rubric scenario are unverified. Documentation is not a successful product test.
  • Usage & Stats is on by default and can be disabled; tokens and cost exist only for turns recorded after the stats database began storing them.
13 atomic answers
  • Logs [supported] History, Director's Notes and the Cue activity log record agent activity. Sources: maestro:nov21-history, maestro:nov21-director, maestro:nov21-cue.
  • Traces [unknown] Traces: The November first-party documentation and repository pass did not establish truthful status and operational statistics at the rubric's required scope. No installed-product fixture was run.
  • Failure reasons [partial] Resilience outages are recorded as recovered or stopped and exported; goal-run stop reasons go to History. Sources: maestro:nov21-usage, maestro:nov21-resilience, maestro:nov21-autorun.
  • Usage export [supported] JSON or CSV export of the selected range includes query events with per-turn tokens and cost, Auto Run runs and tasks, resilience outages, and token use by agent, model, project and account. Sources: maestro:ade154-usage, maestro:nov21-usage.
  • Task stats [supported] Auto Run tab: runs, unfinished runs, tasks done out of attempted and success rate. Sources: maestro:nov21-usage.
  • Agent stats [supported] Agents tab cards show each agent's queries, open tabs and Auto Run share; the agent detail adds tokens and cost. Sources: maestro:ade154-usage.
  • Project stats [partial] Group tiles roll up member agents' queries, time, tokens and cost; token use by project is exported. Sources: maestro:nov21-usage.
  • Time stats [supported] Activity and duration trends over selectable time ranges. Sources: maestro:ade154-usage.
  • Failure status [unknown] Failure status: The November first-party documentation and repository pass did not establish truthful status and operational statistics at the rubric's required scope. No installed-product fixture was run.
  • Approval status [unknown] Approval status: The November first-party documentation and repository pass did not establish truthful status and operational statistics at the rubric's required scope. No installed-product fixture was run.
  • Reconnect consistency [unknown] Reconnect consistency: The November first-party documentation and repository pass did not establish truthful status and operational statistics at the rubric's required scope. No installed-product fixture was run.
  • Performance method [unknown] Performance method: The November first-party documentation and repository pass did not establish truthful status and operational statistics at the rubric's required scope. No installed-product fixture was run.
  • Responsiveness [unknown] Responsiveness: The November first-party documentation and repository pass did not establish truthful status and operational statistics at the rubric's required scope. No installed-product fixture was run.

Sources: maestro:nov21-usage, maestro:nov21-resilience, maestro:nov21-autorun, maestro:nov21-history, maestro:nov21-director, maestro:nov21-cue, maestro:ade154-usage

No sourced public finding not-published

No publishable sourced finding is available for this criterion. This does not mean the product lacks the capability.

Agent cross-checked partial

Group Chat queues busy agents and exposes explicit queue pause/resume. Resource-aware scheduling remains unknown.

Scope
Surfaces: desktop, ssh-remote · Version: First-party documentation/source checked 2026-11-07; exact pinned source revision is recorded in each citation. No installed-product fixture. · Plan: Desktop local/SSH; headless Cue is separately documented with unresolved release parity · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
unknown
Delivery
unknown
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • Only the named atomic facts are established by documentation. Remaining rubric scenarios and unknown answers were not demonstrated.
12 atomic answers
  • Concurrency [partial] Cue: 1-10 concurrent runs per agent with a bounded event queue; Group Chat queues busy agents. No CPU or RAM awareness. Sources: maestro:nov21-cue-config, maestro:nov21-groupchat.
  • Queues [supported] Chat message queues persist across app quit and reopen paused; failed sends remain queued. Sources: maestro:gap-group.
  • Cpu ram [unknown] Cpu ram: The November first-party documentation and repository pass did not establish compute-aware scheduling and verification at the rubric's required scope. No installed-product fixture was run.
  • Backpressure [supported] Busy participant requests wait by default, up to 15 minutes. Sources: maestro:gap-group.
  • Host awareness [unknown] Host awareness: The November first-party documentation and repository pass did not establish compute-aware scheduling and verification at the rubric's required scope. No installed-product fixture was run.
  • Check reuse [unknown] Check reuse: The November first-party documentation and repository pass did not establish compute-aware scheduling and verification at the rubric's required scope. No installed-product fixture was run.
  • Check deduplication [unknown] Check deduplication: The November first-party documentation and repository pass did not establish compute-aware scheduling and verification at the rubric's required scope. No installed-product fixture was run.
  • Bounded tests [unknown] Bounded tests: The November first-party documentation and repository pass did not establish compute-aware scheduling and verification at the rubric's required scope. No installed-product fixture was run.
  • Cancellation [supported] Stop All pauses the queue so queued work does not immediately restart the room. Sources: maestro:gap-group.
  • Fairness [unknown] Fairness: The November first-party documentation and repository pass did not establish compute-aware scheduling and verification at the rubric's required scope. No installed-product fixture was run.
  • Recovery [unknown] Recovery: The November first-party documentation and repository pass did not establish compute-aware scheduling and verification at the rubric's required scope. No installed-product fixture was run.
  • Shared limits [unknown] Shared limits: The November first-party documentation and repository pass did not establish compute-aware scheduling and verification at the rubric's required scope. No installed-product fixture was run.

Sources: maestro:gap-group, maestro:nov21-cue-config, maestro:nov21-groupchat

Human collaboration and control

Agent cross-checked partial

Multiple people can prompt the same Maestro through browser clients. Optional Web Login attributes each browser message to its sender; browser agent/tab selections stay independent.

Scope
Surfaces: ssh-remote · Version: First-party documentation/source checked 2026-11-07; exact pinned source revision is recorded in each citation. No installed-product fixture. · Plan: Desktop local/SSH; headless Cue is separately documented with unresolved release parity · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • Only the named atomic answers are established by this source pass. Remaining atomic answers retain their documented gaps.
11 atomic answers
  • Visibility [supported] Per-host shared history files show what teammates did on a shared SSH project. Sources: maestro:nov07-ssh.
  • Discussion [unknown] Discussion: The November first-party documentation and repository pass did not establish live human collaboration at the rubric's required scope. No installed-product fixture was run.
  • Comments [unknown] Comments: The November first-party documentation and repository pass did not establish live human collaboration at the rubric's required scope. No installed-product fixture was run.
  • Presence [unknown] Presence: The November first-party documentation and repository pass did not establish live human collaboration at the rubric's required scope. No installed-product fixture was run.
  • Coprompting [supported] Multiple signed-in browser users can send messages to the same Maestro agent sessions. Sources: maestro:gap39-login-doc.
  • Ordering [unknown] Ordering: The November first-party documentation and repository pass did not establish live human collaboration at the rubric's required scope. No installed-product fixture was run.
  • Conflicts [unknown] Conflicts: The November first-party documentation and repository pass did not establish live human collaboration at the rubric's required scope. No installed-product fixture was run.
  • Attribution [supported] History entries carry a user pill, sender filter and user_name usage column for signed-in browser prompts. Sources: maestro:gap39-login-doc.
  • Permissions [unsupported] No roles: every Web Login account is an equal operator. Sources: maestro:nov21-remote.
  • Plan restrictions [unknown] Plan restrictions: The November first-party documentation and repository pass did not establish live human collaboration at the rubric's required scope. No installed-product fixture was run.
  • Owner online [supported] The desktop app must keep running; it is the single source of truth for browser clients. Sources: maestro:nov21-remote.

Sources: maestro:nov07-ssh, maestro:gap39-login-doc, maestro:nov21-remote

Agent cross-checked partial

The scoped Web Login uses local username/password accounts. Every account is an equal operator, with no organization, project or task roles. Inspection of the web authentication implementation found no SAML, OIDC or other SSO integration.

Scope
Surfaces: desktop · Version: First-party source pass 2026-11-13; pinned code revisions are recorded in each cited source. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-13 (9 days old)
Agent cross-check
2026-11-21
  • Only the named atomic answers are established by this source pass. Remaining atomic answers retain their documented gaps.
9 atomic answers
  • Organization roles [unsupported] All web accounts are equal operators; no organization role system is provided. Sources: maestro:gap39-login-doc, maestro:gap39-login-code.
  • Project roles [unsupported] There are no scoped project roles; logged-in users operate the same desktop. Sources: maestro:gap39-login-doc, maestro:gap39-login-code.
  • Task roles [unsupported] There are no scoped task roles. Sources: maestro:gap39-login-doc, maestro:gap39-login-code.
  • Membership [partial] The desktop adds and removes Web Login accounts; browsers cannot. Sources: maestro:nov21-remote.
  • Sso [unsupported] The inspected web auth store and routes implement local passwords and sessions, with no identity-provider federation. Sources: maestro:gap39-login-doc, maestro:gap39-login-code.
  • Provisioning [unknown] Provisioning: The November first-party documentation and repository pass did not establish roles and access boundaries at the rubric's required scope. No installed-product fixture was run.
  • Unauthorized access [unknown] Unauthorized access: The November first-party documentation and repository pass did not establish roles and access boundaries at the rubric's required scope. No installed-product fixture was run.
  • Revocation [unknown] Revocation: The November first-party documentation and repository pass did not establish roles and access boundaries at the rubric's required scope. No installed-product fixture was run.
  • Plan scope [unknown] Plan scope: The November first-party documentation and repository pass did not establish roles and access boundaries at the rubric's required scope. No installed-product fixture was run.

Sources: maestro:gap39-login-doc, maestro:gap39-login-code, maestro:nov21-remote

Agent cross-checked partial

Phone access uses a bearer URL token with full host control.

Scope
Surfaces: desktop, ssh-remote · Version: First-party sources checked 2026-11-07; pinned implementations retain their cited revision. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • The remaining atomic questions and the rubric scenario are unverified. Documentation is not a successful product test.
14 atomic answers
  • Command [unknown] Command: The November first-party documentation and repository pass did not establish approvals and temporary access at the rubric's required scope. No installed-product fixture was run.
  • Tool [partial] Coworking confirm policy (off, dangerous by default, all) gates browser and terminal tool calls; browser_eval always needs approval. Other tool approvals are the provider CLI's. Sources: maestro:nov21-coworking, maestro:nov21-install.
  • Network [unknown] Network: The November first-party documentation and repository pass did not establish approvals and temporary access at the rubric's required scope. No installed-product fixture was run.
  • Secret [unknown] Secret: The November first-party documentation and repository pass did not establish approvals and temporary access at the rubric's required scope. No installed-product fixture was run.
  • Delivery [unknown] Delivery: The November first-party documentation and repository pass did not establish approvals and temporary access at the rubric's required scope. No installed-product fixture was run.
  • Approver identity [unknown] Approver identity: The November first-party documentation and repository pass did not establish approvals and temporary access at the rubric's required scope. No installed-product fixture was run.
  • Lifetime [supported] Disabling LIVE rotates token and returns loopback binding. Sources: maestro:nov07-remote.
  • Allow [unknown] Allow: The November first-party documentation and repository pass did not establish approvals and temporary access at the rubric's required scope. No installed-product fixture was run.
  • Deny [unknown] Deny: The November first-party documentation and repository pass did not establish approvals and temporary access at the rubric's required scope. No installed-product fixture was run.
  • Revoke [unknown] Revoke: The November first-party documentation and repository pass did not establish approvals and temporary access at the rubric's required scope. No installed-product fixture was run.
  • Tunnels [supported] Optional Cloudflare HTTPS tunnel while host is running. Sources: maestro:nov07-remote.
  • Defaults [supported] OFFLINE listens on loopback; LIVE on all interfaces. Sources: maestro:nov07-remote.
  • Safer modes [partial] Token holder has full control; no scoped remote roles established. Sources: maestro:nov07-remote.
  • Enforcement [unknown] Enforcement: The November first-party documentation and repository pass did not establish approvals and temporary access at the rubric's required scope. No installed-product fixture was run.

Sources: maestro:nov07-remote, maestro:nov21-coworking, maestro:nov21-install

Sourced partial

History persists summaries with session links, supports deletion and has a configured buffer.

Scope
Surfaces: desktop, ssh-remote · Version: First-party sources checked 2026-11-07; pinned implementations retain their cited revision. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
Not separately cross-checked
  • The remaining atomic questions and the rubric scenario are unverified. Documentation is not a successful product test.
  • Only the named atomic facts are established by documentation. Remaining rubric scenarios and unknown answers were not demonstrated.
11 atomic answers
  • Actions [supported] AUTO, USER, AGENT and optional CUE entries. Sources: maestro:nov07-history.
  • Approvals [unknown] Approvals: The November first-party documentation and repository pass did not establish decision and activity history at the rubric's required scope. No installed-product fixture was run.
  • Edits [unknown] Edits: The November first-party documentation and repository pass did not establish decision and activity history at the rubric's required scope. No installed-product fixture was run.
  • Discussion [supported] History stores summaries and full response content linked to the agent conversation. Sources: maestro:gap-history.
  • Decisions [unknown] Decisions: The November first-party documentation and repository pass did not establish decision and activity history at the rubric's required scope. No installed-product fixture was run.
  • Cross user persistence [unknown] Cross user persistence: The November first-party documentation and repository pass did not establish decision and activity history at the rubric's required scope. No installed-product fixture was run.
  • Cross session persistence [supported] History can reopen the linked session. Sources: maestro:nov07-history.
  • Export [supported] History is file-backed JSON per session; shared SSH history uses per-hostname JSONL files. Formal portable export guarantee not established. Sources: maestro:gap-history, maestro:nov07-ssh.
  • Mutability [partial] User annotations exist; tamper-resistant retention not established. Sources: maestro:nov07-history.
  • Retention [supported] Maximum Log Buffer defaults to 10000 history entries per session. Sources: maestro:gap-history.
  • Integrity [unknown] Integrity: The November first-party documentation and repository pass did not establish decision and activity history at the rubric's required scope. No installed-product fixture was run.

Sources: maestro:nov07-history, maestro:gap-history

Agent cross-checked partial

Git menus and keyboard shortcuts expose diff and commit context.

Scope
Surfaces: desktop, ssh-remote · Version: First-party sources checked 2026-11-07; pinned implementations retain their cited revision. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-07
  • The remaining atomic questions and the rubric scenario are unverified. Documentation is not a successful product test.
13 atomic answers
  • Pending reviews [unknown] Pending reviews: The November first-party documentation and repository pass did not establish review experience and guidance at the rubric's required scope. No installed-product fixture was run.
  • Change understanding [supported] Side-by-side syntax-highlighted diff. Sources: maestro:nov07-git.
  • Evidence navigation [supported] Commit log entry opens that commit diff. Sources: maestro:nov07-git.
  • Comments [unknown] Comments: The November first-party documentation and repository pass did not establish review experience and guidance at the rubric's required scope. No installed-product fixture was run.
  • Request changes [unknown] Request changes: The November first-party documentation and repository pass did not establish review experience and guidance at the rubric's required scope. No installed-product fixture was run.
  • Shipping guidance [unknown] Shipping guidance: The November first-party documentation and repository pass did not establish review experience and guidance at the rubric's required scope. No installed-product fixture was run.
  • Effort [unknown] Effort: The November first-party documentation and repository pass did not establish review experience and guidance at the rubric's required scope. No installed-product fixture was run.
  • Missed evidence [unknown] Missed evidence: The November first-party documentation and repository pass did not establish review experience and guidance at the rubric's required scope. No installed-product fixture was run.
  • Keyboard [supported] Git log/diff shortcuts and command palette. Sources: maestro:nov07-git.
  • Narrow screen [unknown] Narrow screen: The November first-party documentation and repository pass did not establish review experience and guidance at the rubric's required scope. No installed-product fixture was run.
  • Context switching [unknown] Context switching: The November first-party documentation and repository pass did not establish review experience and guidance at the rubric's required scope. No installed-product fixture was run.
  • Roles [unknown] Roles: The November first-party documentation and repository pass did not establish review experience and guidance at the rubric's required scope. No installed-product fixture was run.
  • Retained context [unknown] Retained context: The November first-party documentation and repository pass did not establish review experience and guidance at the rubric's required scope. No installed-product fixture was run.

Sources: maestro:nov07-git

Agent cross-checked partial

Another person can drive the same agents through Web Login while the desktop runs; switching providers loses history.

Scope
Surfaces: desktop, ssh-remote · Version: Maestro v1.0.4 (508ef12, released 2026-11-21): pinned docs and source rechecked 2026-11-21. No installed-product fixture. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
unknown
Delivery
unknown
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • Only the cited atomic answers are established from documentation or pinned source; remaining atoms stay unknown and no hands-on test was run.
12 atomic answers
  • Ongoing takeover [partial] With Web Login another person can drive the same agents from a browser, attributed per message, while the desktop app keeps running. Sources: maestro:nov21-remote.
  • Same task [unknown] Same task: The November first-party documentation and repository pass did not establish colleague takeover and continuity at the rubric's required scope. No installed-product fixture was run.
  • Session history [unknown] Session history: The November first-party documentation and repository pass did not establish colleague takeover and continuity at the rubric's required scope. No installed-product fixture was run.
  • Decisions [unknown] Decisions: The November first-party documentation and repository pass did not establish colleague takeover and continuity at the rubric's required scope. No installed-product fixture was run.
  • Artifacts [unknown] Artifacts: The November first-party documentation and repository pass did not establish colleague takeover and continuity at the rubric's required scope. No installed-product fixture was run.
  • Permissions [unknown] Permissions: The November first-party documentation and repository pass did not establish colleague takeover and continuity at the rubric's required scope. No installed-product fixture was run.
  • Runtime context [unknown] Runtime context: The November first-party documentation and repository pass did not establish colleague takeover and continuity at the rubric's required scope. No installed-product fixture was run.
  • Owner disconnect [unknown] Owner disconnect: The November first-party documentation and repository pass did not establish colleague takeover and continuity at the rubric's required scope. No installed-product fixture was run.
  • Ownership [unknown] Ownership: The November first-party documentation and repository pass did not establish colleague takeover and continuity at the rubric's required scope. No installed-product fixture was run.
  • Attribution [unknown] Attribution: The November first-party documentation and repository pass did not establish colleague takeover and continuity at the rubric's required scope. No installed-product fixture was run.
  • Context loss [unknown] Context loss: The November first-party documentation and repository pass did not establish colleague takeover and continuity at the rubric's required scope. No installed-product fixture was run.
  • Harness transfer [partial] Switching an agent to another provider drops its history and memories (roadmap); Claude Code accounts can share sessions. Sources: maestro:nov21-multiprov.

Sources: maestro:nov21-multiprov, maestro:nov21-remote

Verification and delivery

Agent cross-checked partial

History links completed Auto Run summaries to sessions and Git tools expose diffs.

Scope
Surfaces: desktop, ssh-remote · Version: First-party sources checked 2026-11-07; pinned implementations retain their cited revision. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • The remaining atomic questions and the rubric scenario are unverified. Documentation is not a successful product test.
13 atomic answers
  • Diffs [supported] Working-tree and commit diff views. Sources: maestro:nov07-git, maestro:nov07-history.
  • Documents [unknown] Documents: The November first-party documentation and repository pass did not establish retained artifacts and outputs at the rubric's required scope. No installed-product fixture was run.
  • Screenshots [partial] Pasted screenshots stay in transcripts, visible on desktop and web; no gallery. Sources: maestro:nov21-remote.
  • Recordings [unknown] Recordings: The November first-party documentation and repository pass did not establish retained artifacts and outputs at the rubric's required scope. No installed-product fixture was run.
  • Previews [unknown] Previews: The November first-party documentation and repository pass did not establish retained artifacts and outputs at the rubric's required scope. No installed-product fixture was run.
  • Test outputs [unknown] Test outputs: The November first-party documentation and repository pass did not establish retained artifacts and outputs at the rubric's required scope. No installed-product fixture was run.
  • Reviewer outputs [unknown] Reviewer outputs: The November first-party documentation and repository pass did not establish retained artifacts and outputs at the rubric's required scope. No installed-product fixture was run.
  • Persistence [partial] History stores summaries/full response content; full artifact-gallery semantics unverified. Sources: maestro:nov07-history.
  • Child findings [unknown] Child findings: The November first-party documentation and repository pass did not establish retained artifacts and outputs at the rubric's required scope. No installed-product fixture was run.
  • Prompt leakage [unknown] Prompt leakage: The November first-party documentation and repository pass did not establish retained artifacts and outputs at the rubric's required scope. No installed-product fixture was run.

Sources: maestro:nov07-git, maestro:nov07-history, maestro:nov21-remote

Sourced partial

Auto Run has human-gate and halt markers; History Validated is explicitly only a bookmark.

Scope
Surfaces: desktop, ssh-remote · Version: First-party documentation/source checked 2026-11-07; exact pinned source revision is recorded in each citation. No installed-product fixture. · Plan: Desktop local/SSH; headless Cue is separately documented with unresolved release parity · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
unknown
Delivery
unknown
Checked
2026-11-07 (15 days old)
Agent cross-check
Not separately cross-checked
  • Only the named atomic facts are established by documentation. Remaining rubric scenarios and unknown answers were not demonstrated.
11 atomic answers
  • Tests [unknown] Tests: The November first-party documentation and repository pass did not establish verification and spec gates at the rubric's required scope. No installed-product fixture was run.
  • Lint [unknown] Lint: The November first-party documentation and repository pass did not establish verification and spec gates at the rubric's required scope. No installed-product fixture was run.
  • Types [unknown] Types: The November first-party documentation and repository pass did not establish verification and spec gates at the rubric's required scope. No installed-product fixture was run.
  • Ci [unknown] Ci: The November first-party documentation and repository pass did not establish verification and spec gates at the rubric's required scope. No installed-product fixture was run.
  • Independent review [unknown] Independent review: The November first-party documentation and repository pass did not establish verification and spec gates at the rubric's required scope. No installed-product fixture was run.
  • Spec checks [unknown] Spec checks: The November first-party documentation and repository pass did not establish verification and spec gates at the rubric's required scope. No installed-product fixture was run.
  • Required checks [unknown] Required checks: The November first-party documentation and repository pass did not establish verification and spec gates at the rubric's required scope. No installed-product fixture was run.
  • Failure blocking [supported] Halt marker stops the full playbook; ordinary task failure need not invalidate later independent tasks. Sources: maestro:gap-auto.
  • Requirement conflicts [unknown] Requirement conflicts: The November first-party documentation and repository pass did not establish verification and spec gates at the rubric's required scope. No installed-product fixture was run.
  • Configured policy [supported] HITL markers require desktop human acknowledgement before that task proceeds; CLI skips gated documents. Sources: maestro:gap-auto.
  • Enforced policy [unknown] Enforced policy: The November first-party documentation and repository pass did not establish verification and spec gates at the rubric's required scope. No installed-product fixture was run.

Sources: maestro:gap-auto, maestro:gap-history

Agent cross-checked partial

Git menu integrates worktrees, branch changes, push/pull and PR creation.

Scope
Surfaces: desktop, ssh-remote · Version: First-party sources checked 2026-11-07; pinned implementations retain their cited revision. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-07
  • The remaining atomic questions and the rubric scenario are unverified. Documentation is not a successful product test.
9 atomic answers
  • Branches [supported] Branch picker. Sources: maestro:nov07-git.
  • Worktrees [supported] Nested worktree agents. Sources: maestro:nov07-git.
  • Commits [unknown] Commits: The November first-party documentation and repository pass did not establish git and pull request workflow at the rubric's required scope. No installed-product fixture was run.
  • Pull requests [supported] Create Pull Request action. Sources: maestro:nov07-git.
  • Comments [unknown] Comments: The November first-party documentation and repository pass did not establish git and pull request workflow at the rubric's required scope. No installed-product fixture was run.
  • Ci response [unknown] Ci response: The November first-party documentation and repository pass did not establish git and pull request workflow at the rubric's required scope. No installed-product fixture was run.
  • Rebase [unknown] Rebase: The November first-party documentation and repository pass did not establish git and pull request workflow at the rubric's required scope. No installed-product fixture was run.
  • Conflicts [unknown] Conflicts: The November first-party documentation and repository pass did not establish git and pull request workflow at the rubric's required scope. No installed-product fixture was run.
  • Product state [supported] Ahead/behind counts and branch context. Sources: maestro:nov07-git.

Sources: maestro:nov07-git

No sourced public finding not-published

No publishable sourced finding is available for this criterion. This does not mean the product lacks the capability.

Agent cross-checked partial

History keeps task summaries, sessions, duration and cost; worktree runs can open a PR.

Scope
Surfaces: desktop · Version: Maestro v1.0.4 (508ef12, released 2026-11-21): pinned docs and source rechecked 2026-11-21. No installed-product fixture. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
unknown
Delivery
unknown
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • Only the cited atomic answers are established from documentation or pinned source; remaining atoms stay unknown and no hands-on test was run.
9 atomic answers
  • Delivery status [unknown] Delivery status: The November first-party documentation and repository pass did not establish completion evidence and rollback at the rubric's required scope. No installed-product fixture was run.
  • Retained evidence [partial] History keeps each Auto Run task's summary, session ID, duration and cost; worktree runs can open a PR on completion. Sources: maestro:nov21-history, maestro:nov21-autorun.
  • Rollback [unknown] Rollback: The November first-party documentation and repository pass did not establish completion evidence and rollback at the rubric's required scope. No installed-product fixture was run.
  • Revert [unknown] Revert: The November first-party documentation and repository pass did not establish completion evidence and rollback at the rubric's required scope. No installed-product fixture was run.
  • Stale approval invalidation [unknown] Stale approval invalidation: The November first-party documentation and repository pass did not establish completion evidence and rollback at the rubric's required scope. No installed-product fixture was run.

Sources: maestro:nov21-history, maestro:nov21-autorun

No sourced public finding not-published

No publishable sourced finding is available for this criterion. This does not mean the product lacks the capability.

Ownership, privacy and economics

Agent cross-checked partial

Maestro root distribution uses AGPL-3.0. The embedded web server is AGPL with the app; hosted leaderboard and telemetry services have no published source.

Scope
Surfaces: desktop, ssh-remote · Version: First-party sources checked 2026-11-07; pinned implementations retain their cited revision. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • The remaining atomic questions and the rubric scenario are unverified. Documentation is not a successful product test.
6 atomic answers

Sources: maestro:nov07-license, maestro:nov21-hosted-endpoints, maestro:nov21-achievements, maestro:nov21-home, maestro:nov21-install, maestro:nov21-license, maestro:nov21-pwa

Agent cross-checked partial

Maestro runs locally and can export Cue bundles for user-owned headless servers. It runs without a Maestro account; dependencies are provider CLIs, cloudflared, gh and optional telemetry.

Scope
Surfaces: desktop, Maestro Cue server · Version: First-party documentation/source checked 2026-11-07; exact pinned source revision is recorded in each citation. No installed-product fixture. · Plan: Desktop local/SSH; headless Cue is separately documented with unresolved release parity · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
unknown
Delivery
unknown
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • Only the named atomic facts are established by documentation. Remaining rubric scenarios and unknown answers were not demonstrated.
  • Headless Cue is documented online but absent from the pinned release source, as recorded in COORD-04; no released headless parity claim is made.
8 atomic answers
  • Self hosting [supported] Core desktop is user-operated; exported Cue pipeline engine also runs headless on Linux/container. Sources: maestro:gap-server, maestro:gap-pass-through.
  • Vendor dependencies [partial] Needs provider CLIs; remote access uses cloudflared quick tunnels (no Cloudflare account); PRs use gh; optional runmaestro.ai leaderboard and telemetry. Sources: maestro:nov21-install, maestro:nov21-remote, maestro:nov21-worktrees, maestro:nov21-hosted-endpoints.
  • Export [supported] CLI bundle export creates ZIP files for agents/pipelines; normal bundles contain secret names, not values. Sources: maestro:gap-server.
  • Import [supported] CLI bundle import maps exported workspaces on the server; conflicts need explicit force. Sources: maestro:gap-server.
  • Formats [supported] ZIP bundles plus filesystem workspaces and persistent Cue database. Sources: maestro:gap-server.
  • Cancellation [unknown] Cancellation: The November first-party documentation and repository pass did not establish hosting and exit at the rubric's required scope. No installed-product fixture was run.
  • Account independence [supported] No Maestro account is needed to run it; only provider CLIs need logins. Sources: maestro:nov21-install, maestro:nov21-start.
  • History portability [unknown] History portability: The November first-party documentation and repository pass did not establish hosting and exit at the rubric's required scope. No installed-product fixture was run.

Sources: maestro:gap-server, maestro:gap-pass-through, maestro:nov21-install, maestro:nov21-start, maestro:nov21-remote, maestro:nov21-worktrees, maestro:nov21-hosted-endpoints

Agent cross-checked partial

Anonymous check-in and optional folder sync have explicit data and conflict limits.

Scope
Surfaces: desktop, ssh-remote · Version: First-party sources checked 2026-11-07; pinned implementations retain their cited revision. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-10
  • The remaining atomic questions and the rubric scenario are unverified. Documentation is not a successful product test.
  • Only the named atomic facts are established by documentation. Remaining rubric scenarios and unknown answers were not demonstrated.
  • Cross-device sync is labelled Beta, supports one device at a time and has no conflict resolution.
15 atomic answers
  • Code data [supported] Anonymous check-in excludes file paths, project names and agent output. Sources: maestro:gap-config.
  • Prompts [supported] Check-in excludes prompts and agent output. Sources: maestro:nov07-config.
  • Credentials [unknown] Credentials: The November first-party documentation and repository pass did not establish data handling and telemetry at the rubric's required scope. No installed-product fixture was run.
  • Logs [supported] Optional cloud-folder sync includes session/settings data; it is separate from the anonymous check-in. Sources: maestro:gap-config.
  • Artifacts [unknown] Artifacts: The November first-party documentation and repository pass did not establish data handling and telemetry at the rubric's required scope. No installed-product fixture was run.
  • Retention [unknown] Retention: The November first-party documentation and repository pass did not establish data handling and telemetry at the rubric's required scope. No installed-product fixture was run.
  • Deletion [unknown] Deletion: The November first-party documentation and repository pass did not establish data handling and telemetry at the rubric's required scope. No installed-product fixture was run.
  • Training [unknown] Training: The November first-party documentation and repository pass did not establish data handling and telemetry at the rubric's required scope. No installed-product fixture was run.
  • Residency [unknown] Residency: The November first-party documentation and repository pass did not establish data handling and telemetry at the rubric's required scope. No installed-product fixture was run.
  • Subprocessors [unknown] Subprocessors: The November first-party documentation and repository pass did not establish data handling and telemetry at the rubric's required scope. No installed-product fixture was run.
  • Telemetry disclosure [supported] Install UUID, version, platform, arch and theme. Sources: maestro:nov07-config.
  • Telemetry defaults [partial] Launch/daily check-in with update checking; first-run consent was not observed. Sources: maestro:nov07-config.
  • Telemetry controls [supported] Disable Check for updates on startup to disable check-in. Sources: maestro:nov07-config.
  • Provider separation [supported] Maestro passes prompts to provider CLIs; MCP tools, skills, permissions and authentication configured in Claude Code, Codex or OpenCode apply unchanged. Sources: maestro:gap-pass-through.

Sources: maestro:nov07-config, maestro:gap-config

Agent cross-checked partial

Maestro is AGPL-3.0 open source and downloaded from GitHub Releases, with no price or paid plan published; model usage runs on your own provider accounts and plan quotas (e.g. Claude Max). Provider API keys can also fund model usage directly.

Scope
Surfaces: desktop, ssh-remote, Maestro Cue server · Version: First-party pricing sources read 2026-11-10 · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-10 (12 days old)
Agent cross-check
2026-11-21
  • Only the named atomic answers are established by this source pass. Remaining atomic answers retain their documented gaps.
12 atomic answers
  • Currency [unknown] No first-party source states currency for a paid plan. Checked: runmaestro.ai, docs.runmaestro.ai, docs llms.txt index, RunMaestro/Maestro README and LICENSE (AGPL-3.0 does not state free of charge). No pricing page or paid plan was found, but that does not prove none exists.
  • Interval [unknown] No first-party source states interval for a paid plan. Checked: runmaestro.ai, docs.runmaestro.ai, docs llms.txt index, RunMaestro/Maestro README and LICENSE (AGPL-3.0 does not state free of charge). No pricing page or paid plan was found, but that does not prove none exists.
  • Seats [unsupported] No seat licensing: v1.0.4 has no account, license-key or seat check, and the site has no plans. Sources: maestro:nov21-billing-search, maestro:nov21-home.
  • Seat minimum [unknown] No first-party source states seat minimum for a paid plan. Checked: runmaestro.ai, docs.runmaestro.ai, docs llms.txt index, RunMaestro/Maestro README and LICENSE (AGPL-3.0 does not state free of charge). No pricing page or paid plan was found, but that does not prove none exists.
  • Workspace fees [unknown] No first-party source states workspace fees for a paid plan. Checked: runmaestro.ai, docs.runmaestro.ai, docs llms.txt index, RunMaestro/Maestro README and LICENSE (AGPL-3.0 does not state free of charge). No pricing page or paid plan was found, but that does not prove none exists.
  • Included usage [unknown] No first-party source states included usage for a paid plan. Checked: runmaestro.ai, docs.runmaestro.ai, docs llms.txt index, RunMaestro/Maestro README and LICENSE (AGPL-3.0 does not state free of charge). No pricing page or paid plan was found, but that does not prove none exists.
  • Credits [unknown] No first-party source states credits for a paid plan. Checked: runmaestro.ai, docs.runmaestro.ai, docs llms.txt index, RunMaestro/Maestro README and LICENSE (AGPL-3.0 does not state free of charge). No pricing page or paid plan was found, but that does not prove none exists.
  • Trial [unknown] No first-party source states trial for a paid plan. Checked: runmaestro.ai, docs.runmaestro.ai, docs llms.txt index, RunMaestro/Maestro README and LICENSE (AGPL-3.0 does not state free of charge). No pricing page or paid plan was found, but that does not prove none exists.
  • Overages [unknown] No first-party source states overages for a paid plan. Checked: runmaestro.ai, docs.runmaestro.ai, docs llms.txt index, RunMaestro/Maestro README and LICENSE (AGPL-3.0 does not state free of charge). No pricing page or paid plan was found, but that does not prove none exists.
  • Taxes [unknown] No first-party source states taxes for a paid plan. Checked: runmaestro.ai, docs.runmaestro.ai, docs llms.txt index, RunMaestro/Maestro README and LICENSE (AGPL-3.0 does not state free of charge). No pricing page or paid plan was found, but that does not prove none exists.
  • Model cost [supported] The user pays through provider subscriptions or provider API credentials. Sources: maestro:gap39-provider-billing.
  • Compute cost [partial] Agents run on the user's machine or user-configured SSH hosts; Maestro supplies no compute, so compute is paid to whoever provides those machines. Sources: maestro:nov21-ssh, maestro:nov21-install.

Sources: maestro:docs-index-nov10, maestro:home-nov10, maestro:gap39-provider-billing, maestro:nov21-ssh, maestro:nov21-install, maestro:nov21-billing-search, maestro:nov21-home

Agent cross-checked partial

Maestro adds no fixed fee; scenario costs are provider plans or API usage and the user's own compute.

Scope
Surfaces: desktop · Version: Maestro v1.0.4 (508ef12, released 2026-11-21): pinned docs and source rechecked 2026-11-21. No installed-product fixture. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
unknown
Delivery
unknown
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • Only the cited atomic answers are established from documentation or pinned source; remaining atoms stay unknown and no hands-on test was run.
11 atomic answers
  • One person [unknown] One person: The November first-party documentation and repository pass did not establish scenario cost and controls at the rubric's required scope. No installed-product fixture was run.
  • Two people [unknown] Two people: The November first-party documentation and repository pass did not establish scenario cost and controls at the rubric's required scope. No installed-product fixture was run.
  • Five people [unknown] Five people: The November first-party documentation and repository pass did not establish scenario cost and controls at the rubric's required scope. No installed-product fixture was run.
  • Task assumptions [unknown] Task assumptions: The November first-party documentation and repository pass did not establish scenario cost and controls at the rubric's required scope. No installed-product fixture was run.
  • Concurrency [unknown] Concurrency: The November first-party documentation and repository pass did not establish scenario cost and controls at the rubric's required scope. No installed-product fixture was run.
  • Tokens [unknown] Tokens: The November first-party documentation and repository pass did not establish scenario cost and controls at the rubric's required scope. No installed-product fixture was run.
  • Compute hours [unknown] Compute hours: The November first-party documentation and repository pass did not establish scenario cost and controls at the rubric's required scope. No installed-product fixture was run.
  • Fixed cost [supported] No Maestro fee: the app is a free download with no billing code; scenario costs are provider subscriptions or API usage plus the user's own machines. Sources: maestro:nov21-billing-search, maestro:nov21-home, maestro:nov21-install.
  • Variable cost [unknown] Variable cost: The November first-party documentation and repository pass did not establish scenario cost and controls at the rubric's required scope. No installed-product fixture was run.
  • Included cost [unknown] Included cost: The November first-party documentation and repository pass did not establish scenario cost and controls at the rubric's required scope. No installed-product fixture was run.
  • Unknown cost [unknown] Unknown cost: The November first-party documentation and repository pass did not establish scenario cost and controls at the rubric's required scope. No installed-product fixture was run.

Sources: maestro:nov21-billing-search, maestro:nov21-home, maestro:nov21-install

Agent cross-checked partial

Maestro has public issue support and released/RC scopes that differ. Downloads are public, with releases through v1.0.4 on 21 November 2026.

Scope
Surfaces: desktop, ssh-remote · Version: First-party documentation/source checked 2026-11-07; exact pinned source revision is recorded in each citation. No installed-product fixture. · Plan: Desktop local/SSH; headless Cue is separately documented with unresolved release parity · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
unknown
Delivery
unknown
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • Only the named atomic facts are established by documentation. Remaining rubric scenarios and unknown answers were not demonstrated.
8 atomic answers
  • Release stage [supported] Desktop releases are published on the public GitHub Releases page and building from source is documented; several provider integrations are labelled beta. Sources: maestro:gap-pass-through.
  • Access restrictions [supported] Public downloads from GitHub Releases for macOS, Windows and Linux; no sign-up. Sources: maestro:nov21-install, maestro:nov21-releases.
  • Support [supported] Public repository issue tracker and Discord community. Sources: maestro:gap-pass-through.
  • Response commitments [unknown] Response commitments: The November first-party documentation and repository pass did not establish availability and support at the rubric's required scope. No installed-product fixture was run.
  • Release activity [supported] v1.0.1 (Oct 20), v1.0.2 (Oct 21), v1.0.3 (Nov 12) and v1.0.4 (Nov 21, 2026), plus a v1.1.0-RC pre-release. Sources: maestro:nov21-releases.
  • Status history [unknown] Status history: The November first-party documentation and repository pass did not establish availability and support at the rubric's required scope. No installed-product fixture was run.
  • Deprecation [unknown] Deprecation: The November first-party documentation and repository pass did not establish availability and support at the rubric's required scope. No installed-product fixture was run.
  • Export options [unknown] Export options: The November first-party documentation and repository pass did not establish availability and support at the rubric's required scope. No installed-product fixture was run.

Sources: maestro:gap-pass-through, maestro:nov21-install, maestro:nov21-releases

No sourced public finding not-published

No publishable sourced finding is available for this criterion. This does not mean the product lacks the capability.

Agent cross-checked partial

Maestro is a free AGPL-3.0 desktop download from GitHub Releases for macOS, Windows and Linux (v1.0.4, a full release), and the website, README and licence show no paid tier. Model usage is paid separately through the user's own provider accounts.

Scope
Surfaces: desktop · Version: v1.0.4 (508ef1201846f16fdf79f3b75ce17db838aa2dbc, released 2026-11-21) · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
generally-available
Delivery
native
Checked
2026-11-10 (12 days old)
Agent cross-check
2026-11-21
  • No installed-product test was run.
  • No-card and no-payment answers rest on the documented download path and a bounded v1.0.4 source search, not on a vendor statement that Maestro is free of charge.
10 atomic answers
  • Ongoing free [supported] The whole desktop app is a free download under AGPL-3.0, with no paid tier on the website or in the README. Sources: maestro:ade154-release-v104, maestro:nov21-home, maestro:ade154-readme-license.
  • Trial [unknown] Trial: The November first-party documentation and repository pass did not establish useful free tier and evaluation access at the rubric's required scope. No installed-product fixture was run.
  • Credits trial [unknown] Credits trial: The November first-party documentation and repository pass did not establish useful free tier and evaluation access at the rubric's required scope. No installed-product fixture was run.
  • Self host cost [supported] No ADE fee is published; users pay their own model providers and supply any SSH hosts. Sources: maestro:nov21-home, maestro:nov21-billing-search.
  • Paid only [unsupported] Not paid-only: v1.0.4 installers are free downloads and the source has no billing, license-key or paywall check. Sources: maestro:ade154-release-v104, maestro:nov21-billing-search.
  • Early access [unsupported] No gated access: v1.0.4 is a public full release (not a pre-release) on GitHub Releases, and downloads need no account. Sources: maestro:ade154-release-v104, maestro:nov21-install.
  • Free scenarios [unknown] Free scenarios: The November first-party documentation and repository pass did not establish useful free tier and evaluation access at the rubric's required scope. No installed-product fixture was run.
  • Card required [unsupported] No card: downloads need no account or payment. Sources: maestro:nov21-install, maestro:nov21-billing-search.
  • Usage caps [unknown] Usage caps: The November first-party documentation and repository pass did not establish useful free tier and evaluation access at the rubric's required scope. No installed-product fixture was run.
  • Paid evaluation [unsupported] Evaluating Maestro needs no payment; provider accounts may cost money. Sources: maestro:nov21-billing-search, maestro:nov21-home.

Sources: maestro:nov21-install, maestro:nov21-billing-search, maestro:nov21-home, maestro:ade154-release-v104, maestro:ade154-readme-license

Agent cross-checked partial

Maestro has no price to disclose; model costs are the providers'.

Scope
Surfaces: desktop · Version: Maestro v1.0.4 (508ef12, released 2026-11-21): pinned docs and source rechecked 2026-11-21. No installed-product fixture. · Plan: Not specified · OS: Not specified · Execution modes: Not specified · Region: Not specified · Harness: Not specified
Availability
unknown
Delivery
unknown
Checked
2026-11-07 (15 days old)
Agent cross-check
2026-11-21
  • Only the cited atomic answers are established from documentation or pinned source; remaining atoms stay unknown and no hands-on test was run.
9 atomic answers
  • Units [unknown] Units: The November first-party documentation and repository pass did not establish pricing transparency at the rubric's required scope. No installed-product fixture was run.
  • Included limits [unknown] Included limits: The November first-party documentation and repository pass did not establish pricing transparency at the rubric's required scope. No installed-product fixture was run.
  • Overages [unknown] Overages: The November first-party documentation and repository pass did not establish pricing transparency at the rubric's required scope. No installed-product fixture was run.
  • Compute markup [unknown] Compute markup: The November first-party documentation and repository pass did not establish pricing transparency at the rubric's required scope. No installed-product fixture was run.
  • Model markup [unknown] Model markup: The November first-party documentation and repository pass did not establish pricing transparency at the rubric's required scope. No installed-product fixture was run.
  • Minimum commitment [unknown] Minimum commitment: The November first-party documentation and repository pass did not establish pricing transparency at the rubric's required scope. No installed-product fixture was run.
  • Cancellation [unknown] Cancellation: The November first-party documentation and repository pass did not establish pricing transparency at the rubric's required scope. No installed-product fixture was run.
  • Sales only unknowns [unknown] Sales only unknowns: The November first-party documentation and repository pass did not establish pricing transparency at the rubric's required scope. No installed-product fixture was run.
  • Transparency classification [partial] Not applicable as a price list: Maestro publishes no price and has no billing; provider costs are external, and the Usage Dashboard shows cost where providers report it. Sources: maestro:nov21-home, maestro:nov21-billing-search, maestro:nov21-usage.

Sources: maestro:nov21-home, maestro:nov21-billing-search, maestro:nov21-usage

Sources

maestro:nov07-readme

Maestro README

Cross-platform Electron workspace with parallel agent sessions, Auto Run Markdown checklists or goal loops, remote control, usage views and Git worktree tools. Feature docs may include prerelease behavior.

Kind
documented
Locator
Overview, setup, features and component boundaries
Source date
Not stated
Retrieved
2026-11-07
maestro:nov07-cli

Maestro Maestro CLI

CLI sends/resumes agent sessions, runs playbooks and exposes structured JSON and failure codes. Some commands require the app reachable.

Kind
documented
Locator
Sending messages; Global Flags; Auto Run
Source date
Not stated
Retrieved
2026-11-07
maestro:nov07-remote

Maestro Remote control

A browser/phone client uses the full host UI. URL token grants full control; LIVE-off returns loopback-only and rotates the token.

Kind
documented
Locator
Mobile Web Interface; Network Exposure
Source date
Not stated
Retrieved
2026-11-07
maestro:gap39-clients

Maestro distribution targets

The scoped distribution builds Electron desktop packages for macOS, Windows and Linux. The phone client is the built-in browser interface; no native iOS or Android package exists in the inspected build targets.

Kind
source-inspected
Locator
Electron build mac, win and linux targets; docs/remote-control.md Mobile Web Interface
Source date
Not stated
Retrieved
2026-11-13
Commit
1d462eb41f69e01bf848156e3c47997ad971c462
Path
package.json
maestro:nov21-remote

Maestro v1.0.4 docs: Remote Control

The web interface is the full Maestro app served by the desktop app (phone layout on phones) with the same agents, tabs, transcripts including pasted screenshots, Auto Run and settings, live-synced with the desktop window; the desktop app is the single source of truth and a dropped browser reconnects and reloads. The Cloudflare tunnel needs the cloudflared CLI but no Cloudflare account and lasts while Maestro runs. Web Login adds per-person accounts on the desktop; there are no roles, every account is an equal operator, the desktop is the administrator, and a browser can never add, remove or reset an account. Each browser keeps its own active agent and tab.

Kind
documented
Locator
docs/remote-control.md: Mobile Web Interface; Remote Control; Requiring a Login; Connection Handling
Source date
2026-11-21
Retrieved
2026-11-21
maestro:nov21-config-notify

Maestro v1.0.4 docs: Configuration

Settings > Notifications offers OS notifications when an AI task completes, a long-running command finishes or the agent requires attention; a custom command run on task completion (e.g. say, afplay, notify-send); and in-app toasts with configurable duration and position.

Kind
documented
Locator
docs/configuration.md: Notifications & Sound
Source date
2026-11-21
Retrieved
2026-11-21
maestro:nov21-web-notify

Maestro v1.0.4 source: web notifications hook

A web-interface hook requests browser notification permission and shows browser Notification objects for agent_complete, agent_error, autorun_complete, autorun_task_complete and context_warning events received over the open connection. Searching src for PushManager, web-push and VAPID found no Web Push implementation, so browser alerts need the page connected.

Kind
source-inspected
Locator
src/web/hooks/useNotifications.ts: lines 1-40, 232-245
Source date
2026-11-21
Retrieved
2026-11-21
Commit
508ef1201846f16fdf79f3b75ce17db838aa2dbc
Path
src/web/hooks/useNotifications.ts
maestro:nov21-cli

Maestro v1.0.4 docs: Maestro CLI

maestro-cli sends messages to agents, dispatches into desktop tabs, inspects sessions, creates agents and worktree agents, stops, resumes, skips or aborts an Auto Run, sends steering notes, and runs playbooks, run-doc documents and goal runs headlessly by spawning the agent itself whether or not the desktop window is open. Recurring runs from the CLI use the user's own cron.

Kind
documented
Locator
docs/cli.md: Sending Messages; Dispatching to a Desktop Tab; Controlling/Steering a Running Auto Run; run-doc; Scheduling with Cron
Source date
2026-11-21
Retrieved
2026-11-21
maestro:nov21-features

Maestro v1.0.4 docs: Features

Features list custom notification commands on completion, global environment variables applied to all agent processes and terminals (API keys, proxies), session discovery that imports and resumes existing provider sessions, a file explorer with @ file mentions, an image annotator for screenshots sent with prompts, cross-tab message search and custom slash commands.

Kind
documented
Locator
docs/features.md: Custom Notifications; Global Environment Variables; Session Discovery; File Explorer; Image Annotator; Cross-Tab Message Search; Slash Commands
Source date
2026-11-21
Retrieved
2026-11-21
maestro:src2-maestro-release-v100

Maestro v1.0.0 release notes

v1.0.0 release notes say Antigravity CLI, Grok CLI, Hermes, Oh My Pi and Pi joined the providers, running locally or over SSH.

Kind
documented
Locator
GitHub release v1.0.0 notes, providers paragraph
Source date
2026-10-09
Retrieved
2026-11-10
maestro:nov21-collab

Maestro v1.0.4 docs: Agent Collaboration

Cross-agent mentions and Group Chat work with any provider in either direction ("An agent on one provider can consult an agent on another"), including agents over SSH. Group Chat has an appointed moderator that routes the question to participants, follows up and synthesizes; consults land in durable tabs and History entries.

Kind
documented
Locator
docs/agent-collaboration.md: What Both Share; Group Chat
Source date
2026-11-21
Retrieved
2026-11-21
maestro:nov21-groupchat

Maestro v1.0.4 docs: Group Chat

The moderator is chosen from Claude Code, OpenCode or Codex; agents become participants when @mentioned; the moderator routes work via @mentions over several rounds and decides who hears which answer. By default a busy agent is queued rather than given concurrent work.

Kind
documented
Locator
docs/group-chat.md: How It Works; The Moderator's Role; Agent Availability
Source date
2026-11-21
Retrieved
2026-11-21
maestro:nov21-multiprov

Maestro v1.0.4 docs: Multiple Accounts

One agent runs as one account selected by an environment variable: Claude Code CLAUDE_CONFIG_DIR, Codex CODEX_HOME, OpenCode XDG_DATA_HOME; Copilot CLI has no per-agent account selector, and Factory Droid, Grok, Qwen3 Coder, Antigravity, Hermes, Oh My Pi and Pi have none Maestro knows. Local runtimes such as Ollama or LM Studio are used by pointing a provider's base-URL variable at the local server, which must speak that CLI's API. Pointing an agent at a different provider loses its session history and memories; carrying them across is on the roadmap.

Kind
documented
Locator
docs/multi-provider.md: What Works Today; Copilot CLI; Local Models; Switching Providers
Source date
2026-11-21
Retrieved
2026-11-21
maestro:ade154-agent-ids

Maestro v1.0.4 src/shared/agentIds.ts

At v1.0.4 the canonical, closed AGENT_IDS list is terminal, claude-code, codex, gemini-cli, antigravity, qwen3-coder, opencode, factory-droid, hermes, pi, copilot-cli, omp and grok; the comment says a new agent is added by editing this list.

Kind
source-inspected
Locator
AGENT_IDS, lines 16-30, commit 508ef12
Source date
2026-11-21
Retrieved
2026-11-21
Commit
508ef1201846f16fdf79f3b75ce17db838aa2dbc
Path
src/shared/agentIds.ts
maestro:ade154-agent-defs

Maestro v1.0.4 src/main/agents/definitions.ts

Visible provider definitions at v1.0.4: Claude Code, Codex, Antigravity CLI (agy, stream-json output, --conversation resume; --sandbox read-only is not CLI-enforced and no tool-disabling flag exists), Qwen3 Coder, Hermes, Pi, Oh My Pi, OpenCode, Factory Droid, Copilot-CLI and Grok CLI (batch mode only, no interactive PTY). Gemini CLI is hidden, commented "Not shipping; superseded by Antigravity", and the internal terminal agent is hidden.

Kind
source-inspected
Locator
AGENT_DEFINITIONS ids terminal (L181, hidden), gemini-cli (L347-349, hidden), antigravity (L400-440), qwen3-coder (L473), hermes (L514), pi (L547), omp (L582), opencode (L620), factory-droid (L686), copilot-cli (L771), grok (L849); getVisibleAgentDefinitions L971-974
Source date
2026-11-21
Retrieved
2026-11-21
Commit
508ef1201846f16fdf79f3b75ce17db838aa2dbc
Path
src/main/agents/definitions.ts
maestro:ade154-custom-paths

Maestro v1.0.4 src/main/index.ts

At v1.0.4 per-agent settings can override the binary path (customPath) of an agent already in AGENT_IDS; searching src for "custom agent", "customAgent" and "customProvider" finds only that path override and a CLI --provider-path option, with no mechanism to register a new agent type.

Kind
source-inspected
Locator
custom agent paths, lines 1226-1237; search of src for "custom agent", "customAgent", "customProvider"
Source date
2026-11-21
Retrieved
2026-11-21
Commit
508ef1201846f16fdf79f3b75ce17db838aa2dbc
Path
src/main/index.ts
maestro:ade154-installation

Maestro v1.0.4 docs/installation.md

Supported agents at v1.0.4: Claude Code, Codex, OpenCode and Factory Droid (fully integrated); Copilot-CLI, Hermes, Pi, Qwen3 Coder and Oh My Pi (beta integration); Gemini CLI "Planned support". Maestro is a pass-through that orchestrates the conversation flow in batch mode. The list does not name Antigravity CLI or Grok CLI.

Kind
documented
Locator
Requirements, lines 18-32
Source date
2026-11-21
Retrieved
2026-11-21
maestro:ade154-provider-notes

Maestro v1.0.4 docs/provider-notes.md

All providers take custom CLI arguments and environment variables. Model selection: Claude Code via the Anthropic account; Codex, Factory Droid, Hermes and Qwen3 Coder -m; OpenCode --model provider/model; Copilot-CLI --model (multi-model via models.dev); Pi and Oh My Pi supported. Maestro drives providers in batch mode; Claude Code can instead run each turn through its interactive TUI via the bundled maestro-p helper (TUI Wrapper and Dynamic token sources).

Kind
documented
Locator
Custom Configuration; per-provider tables (Claude Code, Codex, OpenCode, Factory Droid, Copilot-CLI, Hermes, Pi, Qwen3 Coder, Oh My Pi); Token Source
Source date
2026-11-21
Retrieved
2026-11-21
maestro:nov07-providers

Maestro Provider notes

Documents CLI transports, credential funding and provider-specific limits, including Standard-mode restrictions over SSH.

Kind
documented
Locator
Token Source; Claude Code; Codex
Source date
Not stated
Retrieved
2026-11-07
maestro:nov21-install

Maestro v1.0.4 docs: Installation

Downloads come from GitHub Releases for macOS, Windows and Linux. Requirements are at least one installed and authenticated provider CLI and optional Git ("for git-aware features"). Maestro is a pass-through: MCP tools, skills, permissions and authentication work as when running the provider directly.

Kind
documented
Locator
docs/installation.md: Download; Requirements
Source date
2026-11-21
Retrieved
2026-11-21
maestro:nov21-start

Maestro v1.0.4 docs: Getting Started

After install, a first run asks four settings questions (typography, theme, update/crash-report switches, CLI install); the user then makes sure a supported provider is installed and authenticated, clicks New Agent, picks the provider and project folder, and starts chatting. No Maestro account step appears.

Kind
documented
Locator
docs/getting-started.md: 1. Install and launch; 2. Create an agent
Source date
2026-11-21
Retrieved
2026-11-21
maestro:nov21-achievements

Maestro v1.0.4 docs: Achievements

An opt-in global leaderboard at runmaestro.ai: users sign up to have Auto Run stats submitted and ranked.

Kind
documented
Locator
docs/achievements.md: Leaderboard
Source date
2026-11-21
Retrieved
2026-11-21
maestro:nov21-home

Maestro homepage

On 2026-11-21 the site offers downloads, documentation, GitHub and an AGPL-3.0 notice with no pricing page, plan, sign-in or credits; it links separate projects including Maestro Observer (beta Claude Code dashboard with its own sign-up) and maestro-p.

Kind
vendor-claimed
Locator
nav, features, footer
Source date
Not stated
Retrieved
2026-11-21
maestro:gap-docs-mcp

Maestro Overview; Available tools; Technical details

Mintlify-hosted public read-only MCP exposes SearchMaestro for documentation search/retrieval, not session/task control. HTTP transport needs no authentication.

Kind
documented
Locator
Overview; Available tools; Technical details
Source date
Not stated
Retrieved
2026-11-07
maestro:gap-pass-through

Maestro README.md

Maestro passes prompts to locally configured provider CLIs and inherits their MCP tools, skills, permissions and authentication. Desktop installation supports user-owned operation. Its documentation MCP server searches docs rather than controlling tasks.

Kind
documented
Locator
How it works; Features; Quick start
Source date
Not stated
Retrieved
2026-11-07
maestro:nov21-slash

Maestro v1.0.4 docs: Slash Commands

Users create custom slash commands whose prompts use template variables such as {{AGENT_NAME}} and {{AGENT_PATH}} substituted at runtime; bundled Spec-Kit and OpenSpec commands are included.

Kind
documented
Locator
docs/slash-commands.md: Template Variables
Source date
2026-11-21
Retrieved
2026-11-21
maestro:nov21-plugins

Maestro v1.0.4 docs: Encore Features & Plugins

Encore Features ship as built-in first-party plugins in Settings > Plugins, listed in the same catalog as community plugins, each with Enable/Disable, its own settings and an itemized permission list. Usage & Stats, Director's Notes, Cue and Symphony start on.

Kind
documented
Locator
docs/encore-features.md: Enabling a feature
Source date
2026-11-21
Retrieved
2026-11-21
maestro:nov21-plugin-arch

Maestro v1.0.4 docs: Plugins architecture

Third-party plugins are folders with a plugin.json manifest under the user data directory, in three trust tiers (data-only, sandboxed code, sandboxed UI) with default-deny capability grants; the system is gated by the plugins Encore flag, off by default.

Kind
documented
Locator
CLAUDE-PLUGINS.md: 30-second mental model; Status / gating
Source date
2026-11-21
Retrieved
2026-11-21
maestro:nov21-coworking

Maestro v1.0.4 docs: Coworking

Coworking lets an agent read terminal scrollback and drive in-app browser tabs through a per-agent MCP server. Each agent has a confirm policy: off, dangerous (default: navigate, eval, type, newTab, closeTab need approval) or all; browser_eval always needs approval.

Kind
documented
Locator
docs/coworking.md: Reading; confirm policy
Source date
2026-11-21
Retrieved
2026-11-21
maestro:nov07-build

Maestro build manifest

Pinned build metadata identifies the implementation used by the documented application.

Kind
source-inspected
Locator
Dependencies, scripts and application metadata
Source date
Not stated
Retrieved
2026-11-07
Commit
afd833d808cc3134ba9a2439f11513faf97c17e5
Path
package.json
maestro:nov21-pwa

Maestro v1.0.4 source: web server static routes

The desktop app's embedded web server serves a PWA manifest and service worker under the token path and wires the manifest into the served page, so phones use the browser web app (installable to a home screen); the repository has no native iOS or Android client.

Kind
source-inspected
Locator
src/main/web-server/routes/staticRoutes.ts: header comment lines 11-12; manifest route ~line 269; service worker route ~line 282
Source date
2026-11-21
Retrieved
2026-11-21
Commit
508ef1201846f16fdf79f3b75ce17db838aa2dbc
Path
src/main/web-server/routes/staticRoutes.ts
maestro:nov21-autorun

Maestro v1.0.4 docs: Auto Run & Playbooks

Auto Run works on markdown checkbox documents in a chosen folder; configurations save as named playbooks; the /wizard Inline Wizard turns a conversation into Auto Run documents. Runs are Spec-Driven (documents worked to completion) or Goal-Driven (free-text goal, exit criteria, iteration limit) and goal runs stop on completion, deadlock, max iterations or stall, recorded in History. Parallel Auto Runs run across agents, and in one repository through worktree sub-agents; a run can be dispatched into a new worktree and open a PR when it completes.

Kind
documented
Locator
docs/autorun-playbooks.md: Setting Up Auto Run; Creating Tasks; Playbooks; Inline Wizard; Goal-Driven Mode; How a Goal Run Stops; Parallel Auto Runs; Run in Worktree
Source date
2026-11-21
Retrieved
2026-11-21
maestro:nov21-cue

Maestro v1.0.4 docs: Maestro Cue

Cue triggers agent prompts from file changes, timers, agent completions, GitHub activity (new PRs, issues, labels) and pending tasks; each subscription targets configured agents, and fan-out fires one trigger against several agents.

Kind
documented
Locator
docs/maestro-cue.md: overview; event types; fan-out
Source date
2026-11-21
Retrieved
2026-11-21
maestro:nov21-symphony

Maestro v1.0.4 docs: Symphony

Symphony lists GitHub issues from registered open-source repositories; starting one creates an agent and branch, and a draft PR on first commit claims the issue. It needs the GitHub CLI.

Kind
documented
Locator
docs/symphony.md: requirements; selecting an issue
Source date
2026-11-21
Retrieved
2026-11-21
maestro:nov07-autorun

Maestro Auto Run and Playbooks

Markdown checklists or goal loops run explicitly; active documents are read-only except paused runs. Per-run model and effort overrides are supported.

Kind
documented
Locator
Setting up; Tasks; Model Override; Human-in-the-loop
Source date
Not stated
Retrieved
2026-11-07
maestro:gap-auto

Maestro docs/autorun-playbooks.md

Goal runs stop on completion, deadlock, finite iteration cap or three stalled iterations. Desktop HITL markers pause until Done, Resume; CLI emits document_gated and skips the document. Three no-progress dispatches abandon a document. Halt markers stop the playbook and prevent replay until removed. Per-run auto-resume defaults to five attempts. Steering notes typed in the Thought Stream are delivered at the start of the next task. Checkboxes and the editor reopen while a run is paused; Reset on Completion keeps working copies in runs/.

Kind
documented
Locator
Creating Tasks; Running Multiple Auto Run Documents; How a Goal Run Stops; Steering a Run in Flight; Auto-Resume After an Error; Human-in-the-Loop Gates; Stalled Documents; Halt Marker
Source date
Not stated
Retrieved
2026-11-07
maestro:nov21-history

Maestro v1.0.4 docs: History

Auto Run completions create History entries with a summary, clickable session ID, duration, cost and timestamp; agent entries name the calling agent and the question and link to the answering conversation; each entry with a session ID has a RESUME button.

Kind
documented
Locator
docs/history.md: Auto Entries; Agent Entries; Entry Details; Resuming Sessions
Source date
2026-11-21
Retrieved
2026-11-21
maestro:nov21-worktrees

Maestro v1.0.4 docs: Git Worktrees

Each worktree agent has its own working directory, branch, conversation history and state. Create Pull Request requires the GitHub CLI (gh) installed and authenticated.

Kind
documented
Locator
docs/git-worktrees.md: git menu; worktree agents; Create Pull Request requirements
Source date
2026-11-21
Retrieved
2026-11-21
maestro:nov07-memories

Maestro Memories

Claude-only project memory viewer reads/edits/deletes Markdown memory; other providers do not expose this panel.

Kind
documented
Locator
Opening; Reading/editing; Finding
Source date
Not stated
Retrieved
2026-11-07
maestro:nov07-config

Maestro Configuration

Update check sends UUID/version/platform/theme telemetry, disabled with update checking. Folder sync warns against simultaneous devices and has no conflict resolution.

Kind
documented
Locator
Anonymous Check-in; Storage; Sync
Source date
Not stated
Retrieved
2026-11-07
maestro:nov21-cue-config

Maestro v1.0.4 docs: Cue Configuration

Global Cue settings: timeout_minutes (default 30) terminates a run, timeout_on_fail break or continue, max_concurrent 1-10 simultaneous Cue runs per agent (default 1) with extra events queued, and queue_size 0-10000 (default 512) beyond which events drop.

Kind
documented
Locator
docs/maestro-cue-configuration.md: settings: timeout_minutes, timeout_on_fail, max_concurrent, queue_size
Source date
2026-11-21
Retrieved
2026-11-21
maestro:gap-group

Maestro docs/group-chat.md

Group Chat moderators coordinate named local/remote agents. By default busy agents wait up to 15 minutes; disabling that safeguard permits concurrent file edits without a lock. Chat queues persist across reload/quit, restart paused, and retain failed sends until manual resume. Group chats can be deleted with their conversation history.

Kind
documented
Locator
Agent availability; Message queue; Remote agents
Source date
Not stated
Retrieved
2026-11-07
maestro:nov07-group

Maestro Group Chat

A moderator coordinates named agents over multiple turns and routes results. Participants do not automatically see each other replies.

Kind
documented
Locator
How it Works; Moderator role; Availability
Source date
Not stated
Retrieved
2026-11-07
maestro:nov07-cue-server

Maestro Cue on a server

Current public docs describe a source-built Linux/systemd or container runner for exported pipelines and agents. It runs unattended in YOLO, with server-side credentials. The same path is absent at the pinned repository revision.

Kind
documented
Locator
Before you start; Build; Install
Source date
Not stated
Retrieved
2026-11-07
maestro:gap-server

Maestro Before you start; Operate; Known limits

Cue can run exported pipelines on Linux/systemd or containers without a desktop or logged-in user. Agents use server-side logins and YOLO permissions. Run database stores activity without output; engine logs hold output. In-flight runs become failed after engine restart. Stopping SSH shell steps stops the SSH client, not the remote command.

Kind
documented
Locator
Before you start; Operate; Known limits
Source date
Not stated
Retrieved
2026-11-07
maestro:nov21-cue-events

Maestro v1.0.4 docs: Cue Event Types

Interval triggers reconcile missed intervals after system sleep. A one-shot subscription whose fire time passes while Maestro is closed fires on next startup if within grace_minutes (default 360); later it expires without firing.

Kind
documented
Locator
docs/maestro-cue-events.md: time.heartbeat behavior; time.once grace_minutes
Source date
2026-11-21
Retrieved
2026-11-21
maestro:nov21-cue-server-rc

Maestro docs: Running Cue on a Server (published site)

On 2026-11-21 the published docs index lists "Running Cue on a Server" (unattended Linux or container runs without the desktop app), but docs/maestro-cue-server.md exists only on the repository rc branch and is absent from the v1.0.4 tag and main, so the released app does not establish it.

Kind
documented
Locator
docs llms.txt index; GitHub contents API for docs/maestro-cue-server.md
Source date
Not stated
Retrieved
2026-11-21
maestro:nov21-resilience

Maestro v1.0.4 docs: Agent Resilience

Retryable provider failures are resent automatically (availability errors back off 30s up to every 30m; quota exhaustion is probed until it clears), including turns from maestro-cli dispatch, Cue pipelines and the web composer; Auto Run batches park and resume. CLI-launched batches do not auto-retry, and pending retries do not survive quitting Maestro.

Kind
documented
Locator
docs/agent-resilience.md: retry table; Auto Run; quitting
Source date
2026-11-21
Retrieved
2026-11-21
maestro:ade154-cue-overview

Maestro v1.0.4 docs/maestro-cue.md

Cue is an event-driven automation engine, an Encore feature on by default, that sends prompts to agents in response to file changes, timers, agent completions, GitHub activity and pending tasks, defined as subscriptions in .maestro/cue.yaml; the engine starts when it finds one.

Kind
documented
Locator
Cue Overview, lines 1-30
Source date
2026-11-21
Retrieved
2026-11-21
maestro:ade154-cue-triggers

Maestro v1.0.4 docs/maestro-cue-events.md

time.heartbeat fires on a periodic timer and repeats at the configured interval; time.scheduled is a cron-like trigger for specific times and weekdays; github.pull_request polls for new PRs through the gh CLI, with github.issue, github.label and webhook.received events in the same file.

Kind
documented
Locator
time.heartbeat (L45-76), time.scheduled (L77-147), github.pull_request (L401), github.issue (L462), github.label (L513), webhook.received (L630)
Source date
2026-11-21
Retrieved
2026-11-21
maestro:ade154-cue-concurrency

Maestro v1.0.4 docs/maestro-cue-advanced.md

max_concurrent limits parallel Cue runs per agent (1-10, default 1, serial); extra events queue up to queue_size (0-10000, default 512) and overflow is dropped with a "Cue queue overflow" toast.

Kind
documented
Locator
Concurrency Control: max_concurrent; queue_size, lines 305-337
Source date
2026-11-21
Retrieved
2026-11-21
maestro:nov07-git

Maestro Git and worktrees

Git log/diff and PR actions integrate with nested worktree agents. Setup scripts run before first agent prompt, on the selected local/SSH host.

Kind
documented
Locator
Git menu; Worktree agents; Setup scripts
Source date
Not stated
Retrieved
2026-11-07
maestro:nov21-director

Maestro v1.0.4 docs: Director's Notes

Director's Notes (a plugin, on by default) aggregates history entries from all agents into one searchable, filterable timeline with AI-generated synopses.

Kind
documented
Locator
docs/director-notes.md: overview
Source date
2026-11-21
Retrieved
2026-11-21
maestro:nov07-ssh

Maestro SSH remote execution

Agents run on user-provided SSH hosts; remote CLI installation and credentials are prerequisites. Optional per-session shared history writes per-hostname JSONL files to the remote project, shown with hostname badges.

Kind
documented
Locator
Overview; Configuring SSH Remotes; Per-Session Configuration; Collaborating over SSH; Limitations
Source date
Not stated
Retrieved
2026-11-07
maestro:gap39-spawn

Maestro process manager

Agent launch selects local PTY/child processes or SSH commands on a user-configured host. The inspected process-manager spawners and runners do not provision a vendor runtime or per-agent containers or VMs. A user may separately choose to run its host in a VM or container.

Kind
source-inspected
Locator
spawn dispatch, lines 224-262; spawners/ChildProcessSpawner.ts lines 380-385; runners/SshCommandRunner.ts
Source date
Not stated
Retrieved
2026-11-13
Commit
1d462eb41f69e01bf848156e3c47997ad971c462
Path
src/main/process-manager/ProcessManager.ts
maestro:nov07-resilience

Maestro Agent resilience

Provider overload/quota errors retry with visible cards; credential/permission/crash failures do not. Pending retries do not survive app quit.

Kind
documented
Locator
Strategies; Never retry; Status card; Quit
Source date
Not stated
Retrieved
2026-11-07
maestro:nov21-usage

Maestro v1.0.4 docs: Usage Dashboard

Group tiles add up queries, time, tokens and cost for every agent in a Left Bar group; the Auto Run tab shows runs, unfinished runs, tasks done out of attempted and success rate; the Resilience panel counts outages recovered vs stopped; JSON/CSV export includes query events, Auto Run runs and tasks, resilience outages and token use by agent, model and project.

Kind
documented
Locator
docs/usage-dashboard.md: Groups; Resilience; Auto Run; Exporting Data; What's Tracked
Source date
2026-11-21
Retrieved
2026-11-21
maestro:ade154-usage

Maestro v1.0.4 docs/usage-dashboard.md

At v1.0.4 the Agents tab shows per-agent queries, tabs and Auto Run share, with tokens and cost in the agent detail; Group tiles add up queries, time, tokens and cost; the Activity tab shows prompts and duration over time across selectable ranges. JSON/CSV export includes query events with per-turn tokens and cost, Auto Run runs and tasks, resilience outages and token use by agent, model, project and account. What's NOT Tracked lists only prompt and response text, file contents and activity outside Maestro (except provider transcripts read by the Tokens tab), so token counts and costs are tracked. Usage & Stats is an Encore feature, on by default, that can be disabled.

Kind
documented
Locator
Agents (L128-176); Groups (L178-200); Activity (L271-297); Time Range (L415); Exporting Data (L451-469); What's Tracked (L504-521); What's NOT Tracked (L523-527); Enabling and Disabling Collection (L535-545)
Source date
2026-11-21
Retrieved
2026-11-21
maestro:gap39-login-doc

Maestro remote control

Multiple people can drive the same Maestro through optional Web Login. Browser prompts have per-user attribution and independent active tabs. The documentation explicitly says there are no roles; every account is an equal operator.

Kind
documented
Locator
Requiring a Login, lines 90-108
Source date
Not stated
Retrieved
2026-11-13
maestro:gap39-login-code

Maestro web login store

The scoped web authentication implementation persists local username/password accounts and browser sessions. Inspection of src/main/web-server/auth and routes/authRoutes.ts for SSO, SAML, OIDC and OAuth found no identity-provider login implementation.

Kind
source-inspected
Locator
WebUserRecord; WebSessionRecord; WebUserStore login/session methods
Source date
Not stated
Retrieved
2026-11-13
Commit
1d462eb41f69e01bf848156e3c47997ad971c462
Path
src/main/web-server/auth/web-user-store.ts
maestro:nov07-history

Maestro History

Timestamped AUTO/USER/AGENT/CUE entries can be searched and reopened by session ID; immutability is not established.

Kind
documented
Locator
Entry types; Filters; Details
Source date
Not stated
Retrieved
2026-11-07
maestro:gap-history

Maestro docs/history.md

History links summaries, session IDs, timestamps and output. Validated is only a personal bookmark, not an execution gate. Entries can be deleted. JSON history defaults to 10000 entries per session; cross-host shared history is enabled per session through SSH Remote Execution.

Kind
documented
Locator
Entry types; Validated flag; Storage; Cross-host history
Source date
Not stated
Retrieved
2026-11-07
maestro:nov07-license

Maestro license

The cited root license governs this repository distribution, subject to component notices.

Kind
source-inspected
Locator
Root license grant
Source date
Not stated
Retrieved
2026-11-07
Commit
afd833d808cc3134ba9a2439f11513faf97c17e5
Path
LICENSE
maestro:nov21-hosted-endpoints

Maestro v1.0.4 source: hosted service endpoints

The app calls runmaestro.ai APIs for the opt-in leaderboard, an anonymous install check-in gated by the update-check preference, and Cue stats telemetry. Only these clients are in the repository; the hosted services' code is not.

Kind
source-inspected
Locator
src/main/ipc/handlers/leaderboard.ts: leaderboard.ts lines 23-24; src/main/checkin.ts lines 1-28; src/main/cue/cue-telemetry.ts line 59
Source date
2026-11-21
Retrieved
2026-11-21
Commit
508ef1201846f16fdf79f3b75ce17db838aa2dbc
Path
src/main/ipc/handlers/leaderboard.ts
maestro:nov21-license

Maestro v1.0.4 source: package manifest and licence

package.json declares "license": "AGPL 3.0" for the single Electron package whose main process includes src/main/web-server; LICENSE is the GNU AGPL v3.

Kind
source-inspected
Locator
package.json: package.json "license" and "main"; LICENSE header
Source date
2026-11-21
Retrieved
2026-11-21
Commit
508ef1201846f16fdf79f3b75ce17db838aa2dbc
Path
package.json
maestro:gap-config

Maestro docs/configuration.md

Update-linked anonymous check-ins exclude prompts, agent output, file paths and project names. Disabling startup update checks disables check-ins. Optional cloud-folder sync includes settings/sessions/groups; simultaneous devices can conflict with last write winning and no conflict resolution.

Kind
documented
Locator
Anonymous check-in; Cross-device sync
Source date
Not stated
Retrieved
2026-11-07
maestro:home-nov10

Maestro homepage

Says Maestro is open source under AGPL-3.0. Downloads come from GitHub Releases, and the page shows no price or pricing page. maestro-p is described as running claude -p on your own Claude Max quota.

Kind
documented
Locator
'The stage door' section; Legal footer (AGPL-3.0 License link); 'Elsewhere' (maestro-p)
Source date
Not stated
Retrieved
2026-11-10
maestro:docs-index-nov10

Maestro docs index

The docs index describes running several accounts of the same provider and spreading work across their plan quotas, so model usage stays on your own provider accounts. No pricing page is listed.

Kind
documented
Locator
Entries 'Multiple Accounts' and 'Agent Resilience'
Source date
Not stated
Retrieved
2026-11-10
maestro:gap39-provider-billing

Maestro provider notes

Claude Code can use the user subscription login or an API key, gateway or cloud-provider credential in the agent environment; API credentials are billed by their provider.

Kind
documented
Locator
Claude Code funding, lines 65-74
Source date
Not stated
Retrieved
2026-11-13
maestro:nov21-ssh

Maestro v1.0.4 docs: SSH Remote Execution

SSH Remote Execution wraps agent commands in SSH on a remote host the user configures, while the local Maestro stays the control centre.

Kind
documented
Locator
docs/ssh-remote-execution.md: Overview; Configuring SSH Remotes
Source date
2026-11-21
Retrieved
2026-11-21
maestro:nov21-releases

Maestro GitHub releases

Releases: v1.0.4 (2026-11-21), v1.0.3 (2026-11-12), v1.0.2 (2026-10-21), v1.0.1 (2026-10-20) and the v1.1.0-RC pre-release (2026-10-19), public on GitHub.

Kind
documented
Locator
release list
Source date
Not stated
Retrieved
2026-11-21
maestro:ade154-release-v104

GitHub release v1.0.4

v1.0.4 was published on 2026-11-21 as a full release (not a pre-release) with free downloadable installers for macOS, Windows and Linux on GitHub Releases.

Kind
documented
Locator
Release assets: maestro-1.0.4-arm64-mac.dmg, maestro-1.0.4-x64-mac.dmg, Maestro-Setup-1.0.4-x64.exe, Maestro-Portable-1.0.4-x64.exe, maestro-1.0.4-x86_64.AppImage, .deb, .rpm (x86_64 and arm64)
Source date
2026-11-21
Retrieved
2026-11-21
maestro:ade154-readme-license

Maestro v1.0.4 README.md

README links the AGPL-3.0 licence and names no paid tier or price.

Kind
documented
Locator
License, lines 209-211
Source date
2026-11-21
Retrieved
2026-11-21

Comparisons with Maestro

Compare Maestro with any other product on adecomparison.com.