---
schema_version: 7
rubric_version: "0.3"
canonical_url: "https://podium.do/compare/agents/harness-by-autonomous-vs-podium/"
data_url: "https://podium.do/compare/agents/harness-by-autonomous-vs-podium/data.json"
generated_at: 2026-11-22
last_checked_at: 2026-11-13
last_cross_checked_at: 2026-11-21
oldest_checked_at: 2026-11-06
publisher: "Podium"
---

# Harness by Autonomous vs Podium

> Podium publishes this catalog and sells an agent development environment. Podium chooses the products, criteria, sources and wording, so treat these pages as vendor-authored research. Check important claims against the linked primary sources. Sourced findings follow the same evidence rules for Podium and its competitors. Facts labeled Podium team research come from Podium's own team and have no linked source. None of this removes the conflict of interest.

> Based on data from adecomparison.com by Podium (https://adecomparison.com; source repository https://github.com/madeinorbit/ade-comparison, snapshot 4322f0c of 2026-11-22), used under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Source links and check dates are kept. Podium arranges the findings into pairs, applies its publication checks, and added its own research where marked "Podium team research". Data license terms: https://github.com/madeinorbit/ade-comparison/blob/4322f0c035f303791413639d1462a138443d8c1e/DATA-LICENSE.md

48 criteria sourced for both products; 37 criteria also agent cross-checked for both. 49 criteria with a sourced finding for at least one product. "Sourced" findings passed publication validation. "Agent cross-checked" findings also received a separate agent source pass, not necessarily a hands-on test. Humans approve the rubric, templates and prototype; they do not review every finding. It does not declare a winner. A missing sourced finding is not evidence that a capability is absent.

Each row gives the finding's status, support, check date and answer. Scope, limitations, atomic answers and sources for every finding are in each product's document, under the criterion ID. Judgment rules are in the rubric.

- Harness by Autonomous source checks: 2026-11-13. Details: https://podium.do/compare/agents/products/harness-by-autonomous/index.md
- Podium source checks: 2026-11-06. Details: https://podium.do/compare/agents/products/podium/index.md
- Rubric: https://podium.do/compare/agents/rubric/index.md
- Canonical HTML: https://podium.do/compare/agents/harness-by-autonomous-vs-podium/
- Machine-readable JSON: https://podium.do/compare/agents/harness-by-autonomous-vs-podium/data.json
- Methodology: https://podium.do/compare/agents/methodology/

## Access and agent choice

### ACC-01: Client access

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Desktop installers cover macOS 12+ (Apple Silicon + Intel), Linux x64/ARM64 and Windows x64/ARM64 with self-update. The same Flutter app runs in the browser (workspace front door, browser pairing, hn view links) and as installed iOS/Android phone apps (live App Store listing; Play builds plus store redirect). The harness CLI, hn TUI and MCP stdio server cover terminal use. Approvals and waiting lists surface in-app, on phones via push/local notices, and on the device. Agents keep working with clients closed and continue from another window, phone or browser.
- **Podium**: Sourced (sourced); support: partial. Checked 2026-11-06 (16 days old); not separately cross-checked. Podium ships desktop, CLI and browser clients, including phone-browser control. The native iPhone app is private beta; the complete download offering contains no native Android app.

### ACC-02: Agent and model choice

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Fourteen named engines run under their own vendor logins, followed via transcripts plus vendor hooks/plugins Harness installs; it never wraps an agent. Local models run through Grid, Ollama, MLX-LM and vLLM, and any pane can retarget to an Autonomous Grid relay. Tabs freely mix engines and machines. No custom engine-adapter mechanism exists.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. Podium launches the native Claude Code, Codex, Grok, OpenCode, Cursor and Pi CLIs in persistent terminals on machines you control. The 0.2.0 release calls Pi fully supported. Models come from each harness catalog, and capabilities differ per harness: Pi declares no quota, handoff or hooks. Mixed-harness operation was not exercised.

### ACC-03: Subscriptions, API keys and vendor credits

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Agent work runs on the user's own engine subscriptions and API-style logins; Harness sells no credits and states no vendor-credit route. Local models and the Grid relay are supported routes. A different account signs in with login --force (one active account). No Harness account or token purchase is needed for local harnesses.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. Podium supports native coding-agent logins and managed provider credentials. Claude subscriptions (via native login or a managed setup-token) and local ChatGPT/Codex login are supported paths, alongside Anthropic and OpenAI API keys. Docs list billing and hosted metering as not implemented product claims.

### ACC-04: Onboarding and project prerequisites

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Download, install, sign in, link machines with a password, then New Harness (machine, folder, engine, advanced). Any folder starts a harness (plain cwd supported; worktrees optional); GitHub URLs clone; per-machine Node 20+ and tmux, macOS 12+ / Ubuntu 22.04+ stated. Headless Linux servers onboard with install.sh. No managed-cloud setup exists.
- **Podium**: Sourced (sourced); support: partial. Checked 2026-11-06 (16 days old); not separately cross-checked. Docs describe a first-run flow (install, podium, register a Git repository, confirm a detected logged-in harness) and a headless VPS installer with machine joining. Setup time and failure rate were not measured, and no self-serve managed cloud setup is documented. The current README and installation guide use podium.do/download and podium.do/install.sh as the canonical stable entrypoints.

### ACC-05: Programmable access

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. A product CLI (harness + hn + harness mcp), a credentialed loopback REST/WebSocket API with named agent/session operations, the same tools as an MCP stdio server and machine-scoped MCP servers cover automation. No public SDK or webhooks exist. The loopback protocol carries a version field; no published stability policy was found.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. The podium CLI covers issue create, show, update, claim, dependencies, artifacts, session send/status/read/stop and review actions, and the server exposes the same issue commands as a bearer-token MCP tool endpoint. No documented public HTTP API, SDK or webhooks were found.

### ACC-06: External tools and extensions

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Machine-scoped MCP servers attach to every agent a machine starts (per-engine/per-project scoping, product-managed); Claude/Codex/Cursor plugins install a Harness skill plus the MCP server; domain harnesses (harness.json folders) bring instructions, tools and viewers; repo conventions (AGENTS.md, agents/<name>.md) feed sessions. Engine-native logins stay inherited from the underlying harness.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-07. Podium gives agents its own issue tools over CLI and MCP, and agent profiles import and export Claude Code agent files. No Podium documentation for user-configured MCP clients, plugins or custom tools was found.

### ACC-07: Client implementation

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Desktop and phone apps are Flutter (desktop targets macOS/Linux/Windows/web; mobile targets Android/iOS); the TUI and installer are Rust; daemon and CLI are managed Node. No Electron shell is used.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. Desktop releases use a Tauri 2 shell that serves the Podium web UI; mobile source uses Expo and React Native. A native iPhone release is described as private beta.

## Planning and context

### PLAN-01: Intake and external tracker relationship

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Local creation (New Harness form, prompts, 1:1 chats) and GitHub-URL cloning are product paths; agents work tasks through Swarm commands. The authoritative tracker is split by design: Swarm tasks for work, GitHub for code and pull requests. Task numbers are stable; PR state is read back and verdicts/merges posted out; issues are not synced. People cannot post in channels or task chats.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. Podium issue records are native and authoritative with stable references, CLI creation and editing, and agent-driven filing from conversation. The in-product assistant has Linear search, create and move tools when a key is configured, but no import or two-way tracker sync was found.

### PLAN-02: Durable task plans

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Swarm tasks are durable server-side goals with chats, sizes, priorities and dependency numbers, linked to the pull requests that close them and surviving restarts. Goal text is writable and messages editable; goal-edit history, explicit acceptance criteria and in-run amendment rules are not established.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-11. Issue brief, acceptance criteria, state paragraph and todos persist on the durable issue record. Release 0.3.0 removed the separate design and notes fields and appended their text to the brief. On the human issue page the title, description and acceptance are editable and the brief is read-only; agents edit the brief and other fields through issue commands.

### PLAN-03: Hierarchy, dependencies and evolving scope

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Swarm projects hold flat task lists with sizes, priorities, dependency numbers and claim limits; ordering is enforced at merge time (lands-after) and by the ready-to-land rule. No epic or parent/child hierarchy exists, and growing a task into an epic is not a construct. No rigid preselected workflow is required.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. Podium supports persistent parent and child issues, phases and typed dependency links, including reparenting an existing issue without replacing its identity. Since 0.3.0 issue types are task, phase and decision: epic and other kinds are retired, and any issue with children acts as a parent. Dependency types are fixed to blocked-by, related, discovered-from, supersedes and duplicates. Cycles are refused and docs list an open blocker as a reason an issue cannot start. No mandatory workflow preset applies in the baseline.

### PLAN-04: Shared context and memory

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Repo AGENTS.md plus per-agent files, DSH instructions and installed skills feed sessions; the Memories agent aggregates every engine's memory dirs with the Harness session index, cited by agent and project, with search. Correction, deletion and freshness controls for memory are not established; takeover continues through SessionStart re-reads and handoff files.
- **Podium**: Sourced (sourced); support: partial. Checked 2026-11-06 (16 days old); not separately cross-checked. v0.3.0 routes Podium issue instructions and relay environment through terminal and engine session launch. Claude stream sessions use appended instructions and boundary callbacks; Codex server sessions use developer instructions, Grok ACP uses rules, and OpenCode uses the prompt system field. Claude and Codex boundary hooks can re-prime after compaction. Transcript mirroring is on by default; full-text indexing needs the command-palette flag, which defaults off. Agent profiles can carry markdown instructions. Curated memory, skills and memory correction or deletion controls remain unestablished. No instruction-delivery or compaction scenario was run.

### PLAN-05: Architecture and API awareness

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Product-level search covers every past and live conversation on a machine; codebase awareness itself is engine-side. Voice input, image paste and craft viewers give multimodal inputs. No product code index exists; retrieval, citation, conflict-detection and enforcement details beyond merge gates are not established.
- **Podium**: Sourced (sourced); support: partial. Checked 2026-11-06 (16 days old); not separately cross-checked. Not established. Podium mirrors daemon transcripts by default, keeps a full-text index only when the command-palette flag is on, and stores session attachments. No repository code search, architecture awareness or API retrieval feature was found.

### PLAN-06: Native issue management and idea backlog

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Swarm keeps persistent tasks (goal+chat) with sizes, priorities, claim limits, dependency numbers, durable open/closed/not-planned states and PM dispatch without accept/decline; tasks and harness sessions are distinct entities. Edits, search, filters, archive and reopen are only partly established, and idea parking was removed in v4. GitHub remains required for code and PRs.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. Podium exposes native issue creation with an optional start flag, separate dispatch, editing, comments, labels, assignment, search, board filters, archive, sibling ordering and an issue tree. Issue records have their own identity separate from sessions and stages are durable. The idea-to-mission scenario was not run. v0.3.0 refuses an agent close or done update while issue sessions are computing unless --confirm-interrupt is supplied; guidance says to request review and close only on user instruction.

### PLAN-07: Agent-native issue operations

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Agents share stable task numbers and agent ids; read/create/update run through documented super-tab verbs, the MCP server and the loopback API; take claims with a limit of 4; results return as posted replies, recaps and cost lines; humans watch with attribution in product views; permissions scope agents to their own sessions and server-enforced reads. Task children, agent-side reassignment and conflict handling are not established.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. CLI and MCP address the same canonical issue records for reads, creation, edits, claims, state, dependencies and artifacts. Relayed agents have worker capability scoped to their issue subtree, with --outside-scope confirmation for other existing issues. attachSession is not subtree-gated, but direct rehoming from a real issue is refused and an agent attach naming a Proposed target is refused. That does not make every Proposed subtree inert: operator attachment or spawn-time cwd binding can leave a session attached to a Proposed issue; newSubissue without targetId then bypasses the proposed-parent guard, creates a human-audience backlog child and attaches the existing session. This path remains at v0.3.0 source and was not exercised live. Concurrent-edit handling remains unknown.

### PLAN-08: Living product specifications

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: unsupported. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. No living product-specification construct exists: no persistent cross-task requirements with human/agent revisions, approvals, implementation links, retrieval, drift detection or review integration. Goals are per-task briefs, AGENTS.md files are instructions, and Memories is a memory store; none is a versioned spec.
- **Podium**: Sourced (sourced); support: partial. Checked 2026-11-06 (16 days old); not separately cross-checked. The spec CLI and service (living specs under pspec with stable SP IDs) are in v0.3.0. Specs is an experimental, default-off feature listed on the edge channel: the flag controls the Specs UI, the spec line in CLI help and whether agents are told about the spec CLI. The podium spec command itself is dispatched, and the server specs service has no flag check, so direct CLI use does not need the flag. This agrees with the docs, which call specs not baseline rather than blocked. Not run end to end.

## Coordination and automation

### COORD-01: Parallel work

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Sessions run concurrently across panes, tabs and machines with working/waiting/idle/failed/stopped states visible in every client; start requirements gate memory, disk and tools (no CPU cap, no count quotas); agent work lists keep at least 3 deep with quota waits and offer queues. Concurrency is documented, not hands-on tested.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. Docs and release notes describe many sessions running as independent terminals, one worktree per issue, and several paired machines under one server, with session and issue state visible on the flight deck, board and CLI. Agent-initiated spawns are limited to 10 per target issue per UTC day. Achieved concurrency is unmeasured.

### COORD-02: Ownership and dispatch

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. The PM agent assigns tasks to agents and agents claim from ordered lists (limit 4) with quota waits and decline reasons; routing prefers the agent that can ship soonest on a machine with memory, disk and tools; ownership is enforced by the identity rule (an agent runs only under its own person's login) with server-enforced reads. Human-to-human assignment does not exist.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. Issues carry an assignee, claim sets assignee and in_progress, a coordinator session can be set per issue, sessions have seats, and work can be placed on a chosen machine or agent profile. claim has no exclusivity check of its own, locks are advisory leases and Podium does not serialize agents editing one worktree, so exclusive ownership is not enforced.

### COORD-03: Inspectable delegation and communication

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. The PM/coordinator slot is product-designated (Cmd-S creates the coordinator; Switch PM moves it) with shipped delegation tools: task assignment, @sender message injection, verdicts, and merge-by-rule, plus Agent Messages tab-to-tab send (experiment, off by default). Live states, retained replies/recaps/handoffs and follow-up via resume round it out; agent-to-agent reassignment is not established.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. Podium orchestrates delegation itself: podium agent spawn starts a full child session on an issue with a chosen harness, model, effort or agent profile, mail carries messages, and bounded await reports the child result. Children are real sessions shown as seats on the flight deck. Agent-initiated spawns take a daily budget per issue. A complete delegated task with retained child findings was not run.

### COORD-04: Unattended triggers

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: unsupported. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. No unattended triggers exist: no schedules, event triggers, recurring tasks, completion conditions, scheduler, overlap handling or retries. Timers in-tree are internal poll/backoff only. Single-shot and conversational initiation is the whole model; Grid-brand schedules belong to the separate Grid product.
- **Podium**: Sourced (sourced); support: partial. Checked 2026-11-06 (16 days old); not separately cross-checked. The server owns a scheduler for cron and one-time automations that spawn agent sessions, skipping overlaps and not backfilling occurrences more than an hour late. The Automations UI is an edge-visibility feature that defaults off and docs list automations as non-baseline. Agents can subscribe to issue and session events.

### COORD-05: Intervention and attention

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Pause, stop, resume, steer (prompts, talk, type), wait-list and pending-decision surfacing, failure reasons and retained transcripts/handoffs are product paths across desktop, phone, device, CLI and API. Escalation beyond waits and the exact chat/terminal parity edge are not established.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. Chat and native CLI views share a session. v0.3.0 defaults human controller chat to steer and immediately types it into running Claude Code, Codex, Grok and OpenCode TUIs that declare native busy-input queuing; undeclared harnesses wait for a boundary, and compaction waits. Mail and automation keep when-ready delivery. Stop, resume, questions, decision issues and offers remain available. This is inspected implementation, not a cross-harness runtime parity result.

### COORD-06: Fleet and epic management UI

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. The Swarm view (channel plus task chats beside coordinator/agent rows), harness swarm CLI reads, roster machine stats and told/pushed attention signals cover fleet visibility; merge ordering and workload gating are enforced in code. Grouping, history navigation and navigation-error coverage are only partly established.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-11. Release 0.3.0 ships a flight deck showing a whole mission in one view: phases, every task and its state, who works on what, what is blocked and by what, and proposals. Done work stays visible. The CLI prints the same snapshot. The waterfall source is dormant and the Timeline view was deleted, so per-task timeline history is not shown. The shared 12-task UX scenario has not been run.

### COORD-07: Workflow flexibility and orchestration visibility

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Cmd-N/Cmd-S free-form creation, natural-language PM direction with voice, per-agent engine/model/profile/permission configuration, in-run engine changes with handoffs and Grid retargets, retained task/PR/verdict links, and Cmd-P/docs/agents-page discoverability with frictionless Google/Apple/QR sign-in. Epic growth and mandatory presets do not exist.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. A quick task can be an interactive session or an issue; an issue can grow into a mission by gaining children while keeping its identity; no mandatory workflow preset applies in the baseline. Per-agent harness, model and effort are chosen at spawn or through agent profiles, and the deck shows seats. The coordinator-with-two-reviewers scenario was not run.

## Execution and reliability

### RUN-01: Laptop-independent execution

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Local execution plus your own linked machines (password-linked, SSH-free) over E2EE relay/WebRTC/TURN, mixed in one tab; the backend, daemon and relay are open source with container files. No Harness-operated compute exists. Agents keep working with clients closed and resume from another window, phone or browser.
- **Podium**: Sourced (sourced); support: partial. Checked 2026-11-06 (16 days old); not separately cross-checked. The published Podium execution choices are local user machines and a user-owned VPS or paired machine. No vendor-operated agent host is in the complete download offering.

### RUN-02: Workspace and process isolation

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Each harness gets its own branch and git worktree with its own tmux pane/process; secrets travel in the tmux environment (never argv/files) with 0600 credentials and redacted handoffs. No container, VM or OS-sandbox isolation exists, and hostile-code containment is not claimed.
- **Podium**: Sourced (sourced); support: partial. Checked 2026-11-06 (16 days old); not separately cross-checked. Podium separates issue files with Git branches and worktrees, but its released daemon launches native host processes without product-managed container or VM provisioning.

### RUN-03: Recovery, expiry and continuity

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Client disconnects, app restarts and machine-offline stretches are designed-for: agents keep running, tabs/state reattach, engines resume where supported, handoff files bridge engine changes in degraded modes, and multi-week Swarm operation is on record. Daemon-crash, backend-restart and duplicate-action details are not established.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. Closing a browser tab does not stop the agent process on a running Podium host. Idle sessions may hibernate under configurable policies; a stopped session keeps its transcript, branch and resume reference. Release 0.2.0 says a failed update rolls back and running sessions keep running through the upgrade; database rollback remains a manual restore. Terminal reattachment goes through the driver adopt contract with process-identity and observation-lease checks. Worker-crash and multi-day behavior were not tested.

### RUN-04: Development environment reliability

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. The daemon installs engines on first launch and per-machine Node/tmux/sqlite prerequisites are documented; loopback port, per-engine homes and 0600 credentials are fixed; remote setup is one install.sh plus link; --repair exists. Caches, services, ports-collision handling, secret injection and install dedupe are not established.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. Only machine setup is documented: the installer installs missing runtime tools and pairs a host, and the state directory holds server-managed secrets. Dependency caching, services, ports and automatic repair for project environments were not found.

### RUN-05: Limits and failure handling

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. No product-side time, turn, token or spend limits exist; token visibility comes from usage windows and ledgers while enforcement stays provider-side (subscription modes) or rule-side (pair dial, merge gates). Provider exits surface as failed states with reasons for manual resume/restart.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. Podium exposes retryable errors, optional auto-continue with backoff, bounded waits and provider quota. Claude stream and hosted headless drivers enforce a supplied turn deadline, with no deadline when timeoutMs is absent. The one-shot headless server runner supplies its budget, default ten minutes. Agent spawns have a daily per-issue budget. Operator-configurable task-wide turn, token and spend caps remain unestablished; quota display is not enforcement.

### RUN-06: Truthful status and operational statistics

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Transcripts, scrollback, device history, journaled automation and one-command log export give logs; start-failure reasons, told-on-PR notes and sha-bound verdicts give failure/approval truth; per-task token rows, scoreboards and usage windows/ledgers give statistics with a documented benchmark method. Traces, usage export and reconnect-consistency details are not established.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-11. Podium offers local logs and crash export, structured session status, issue stats, a usage and analytics sheet with hourly list-price cost, per-task costs and host load panels. The 0.3.0 notes report a vendor benchmark of about 330 MB server memory against about 1.2 GB for an earlier Podium build, and 8 to 52 percent faster start; the method is published in the repository and was not reproduced here. Usage export and cross-reconnect consistency were not established.

### RUN-07: Per-task API and compute cost

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Per-task token counts with hand-reviewed USD list prices post cost lines on landed PRs; per-provider ledgers and subscription windows show allowances; refresh is per minute. Billed charges never exist (BYO subscriptions, never priced), and compute cost, alerts, export and drilldown are not established.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. Since 0.2.0 Podium shows what each task costs, on the task header and the usage sheet, with per-session rows and descendant rollups. The 0.3.0 deck adds a mission cost chip. Figures are token estimates at public model list prices, not subscription invoices or provider bills.

### RUN-08: Compute-aware scheduling and verification

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Starts gate on probed memory/disk/tools with quota flags; agent lists, offer queues, quota waits and the merge train queue work; trains combine checks and hot-file warnings serialize risky landings. CPU has no cap, and backpressure, bounded-test, fairness-beyond-priority and recovery details are only partly established.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. Podium reports host memory, load and disk, auto-hibernates idle sessions under pressure, and provides named advisory resource leases. The merge-queue panel was removed; a shipping queue remains behind the Shipping flag. Concurrency limits and bounded test queuing were not established.

## Human collaboration and control

### TEAM-01: Live human collaboration

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Members share visibility into Swarm work (channel reads, watch/join, told notes) with discussion across channels, task chats, threads and 1:1 agent chats; people comment via GitHub and CLI verdicts since channels are agents-only for posting. Simultaneous co-steering, presence, ordering and plan-tiering do not exist; async progress never needs the owner online.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. The issue UI displays owner and visibility read-only and deliberately omits sharing controls. Backend share handlers and a Members settings section with invites and admin or member roles exist; they do not establish a finished human collaboration workflow. Docs say small teams can share an instance but the interface prioritizes one operator.

### TEAM-02: Roles and access boundaries

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Project-scoped roles (Lead, PM, members, readers/writers) carry documented rules with server enforcement, revocation and unauthorized-access refusals; sign-in is Google/Apple/QR only with no org SSO, SCIM provisioning or organization roles.
- **Podium**: Sourced (sourced); support: partial. Checked 2026-11-06 (16 days old); not separately cross-checked. Podium defines member/admin instance roles and admin-only membership management. The released client authentication uses password-backed sessions and paired-device tokens without enterprise SSO.

### TEAM-03: Approvals and temporary access

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Command approvals ride engine permission modes with phone-side approve/answer, pair autonomy dials (watch/suggest/act-within-rules) and per-launch tokens; the dial and the merge rule permanently refuse push/force/delete/deploy/publish/drop/merge approvals while READ-OK-gated merges still land. Network/secret/tunnel specifics and approval lifetimes are only partly established.
- **Podium**: Sourced (sourced); support: partial. Checked 2026-11-06 (16 days old); not separately cross-checked. Agent-session management requests go through an operator approval broker with approve and deny; the privacy docs say lifecycle commands and automation requests may require operator approval. Harness permission prompts appear as a Permission needed state; credential copying during pairing is a per-grant opt-in. Command-level policy enforcement beyond the harness own modes was not established.

### TEAM-04: Decision and activity history

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Device add/remove history, journaled automation, co-authored commits, signed verdicts and retained chats/decisions give attributed cross-user, cross-session history with message edits and goal writes; project deletion removes its messages. Chat export, edit rules, retention windows and integrity details beyond signed device logs and sha-bound verdicts are not established.
- **Podium**: Sourced (sourced); support: partial. Checked 2026-11-06 (16 days old); not separately cross-checked. Issue actions, approval steps, comments and mail are recorded server-side. Issue activity events, including approval steps, are pruned by the server's janitor once they are older than 14 days or fall outside the newest 50,000 rows. That policy does not define all content retention, export or tamper-evident audit guarantees.

### TEAM-05: Review experience and guidance

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Wait lists, Needs-you badges and PR groupings locate pending reviews; recaps, transcripts, task/PR links and Hub outputs navigate evidence; READ-OK/HOLD plus told notes carry comments, change requests and shipping guidance; task sizes estimate effort; full keyboard remap and the phone app cover narrow screens; context persists across switching.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. Review is a documented loop: issue stage review, diff and checks, artifacts, offers with merge or send-back buttons, and a Tray of review-ready work. Release 0.3.0 adds quoting text from a reply or review finding into the next message and a full-height reading view with an outline. The board has a key map with a help sheet. Guided-review quality and narrow-screen use were not measured.

### TEAM-06: Colleague takeover and continuity

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. A colleague cannot take over a live session (identity rule: an agent runs only under its own person's login), so takeover is by handoff: sealed handoff files with degraded modes, Switch PM ownership transfer, co-authored attribution and continued work after disconnect. Runtime-context transfer limits and decision/artifacts edge details are partial.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. The durable issue, session transcript, artifacts and mail persist for any later agent, and a live session can be handed to another machine in an edge feature. Colleague takeover by a second person was not established.

## Verification and delivery

### SHIP-01: Retained artifacts and outputs

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Typst review ZIPs, viewer captures, Hub portable snapshots with a public gallery, PR-side verdict/merge/cost notes, retained transcripts/handoffs/chats and redacted, never-prompt-carrying handoffs cover artifacts; media previews stream in chunks. Product code-diff galleries, recordings, test-output retention and chat export are not established.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-07. Artifact commands store immutable server-side copies of issue-linked files, including images, video, HTML and documents, which render in the issue sidebar and offers and stay readable after the worktree is gone. Each task has an artifacts page grouped into screens, documents and reports, with filters and search. Revision linking remains untested.

### SHIP-02: Verification and spec gates

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Delivery runs on enforced gates: every check green plus a non-writer READ-OK with no standing HOLD on the exact head, or no merge. The rule, its train and its told notes are product code; spec conformance gates do not exist (no specs), and check suites themselves are the projects' own.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. In the experimental Shipping path a configured validation profile runs before landing (default argv bun run test with a 10 minute timeout) and a failed validation holds the order. Lint, type, CI and spec conformance gates and independent review requirements were not found as product-enforced policy.

### SHIP-03: Git and pull request workflow

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Branch-per-harness and worktree-per-harness are the product model; agents commit and open PRs with the person's login; Harness posts verdicts, merges and told notes, reads CI state, detects conflicts/behind states and links tasks to PRs. Rebase execution and full product-state depth are not established.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. Issues own branches and worktrees, and issue actions rebase, open a pull request with gh (pr create --fill) and merge, subject to a merge-style setting (ff-only by default, with auto-rebase before merge). Podium does not resolve branch conflicts, and PR review comments and CI response were not established.

### SHIP-04: Merge and deployment authority

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Merges are product-executed by rule (train, exact head, person's login) with sha-bound approvals and HOLD vetoes; agents can never merge. Releases and production deploys are not product paths, and pair automation permanently refuses deploy-type approvals.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. The operator accepts work through offers and runs the merge action under a branch merge-lock; in the experimental Shipping path approval is frozen to an exact head and delivery lands on a local target branch by default. Release and deployment are not product features found in the sources.

### SHIP-05: Completion evidence and rollback

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Task-PR co-authored links, cost lines and merge notes give completion evidence with sha-bound approvals that die on new heads; revert titles count against PRs. Build/deployment links, rollback and product build-link depth are not established.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. Shipping orders record approved base and head SHAs, tested integration SHA, landed ref and validation result in a durable delivery receipt, and stale approvals hold the order. Shipping is experimental. The pr action records a PR URL on the issue. Build and deployment links and rollback were not found.

### SHIP-06: Batch review and high-volume shipping

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. The merge train is a dependency-aware delivery queue (lands-after ordering, combined checks, per-head notes, retry-backoff reads) with sha-bound approvals, HOLD change requests, size/effort signals and hot-file warnings; human batch-review UI depth (summaries, grouping, artifact navigation) is not established.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-07. The shipping service records revision-specific approval, configured validation and delivery receipts. The Shipping panel is an experimental edge toggle, off by default; the separate Queues panel was removed.

## Ownership, privacy and economics

### OWN-01: Source and license

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. The whole repo (app, CLI, daemon, relay, backend, device, store) is MIT-licensed public source at the pinned tag; upstream tools keep their own licenses. Hosted backend/relay parity is not established from the outside, and the GitHub App/extension packaging is configuration, not a license exception.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. The public repository is licensed Apache 2.0, and the website presents Podium as an open-source agent development environment. Hosted extensions and hosted-service terms are not established.

### OWN-02: Hosting and exit

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Everything runs on your machines with no account for local work; the MIT sources include backend containers for self-hosting the account/relay side, and Hub snapshots, log export and on-disk transcripts port out while hub/import and repo cloning port in. Account deletion is by email; history portability from the backend is not established.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-11. Local desktop and self-hosted installs run on user-controlled machines with a documented state directory. The docs describe stopped full-state backup and automatic pre-migration database-only snapshots retaining three copies. Those snapshots omit artifacts, transcripts, uploads, configuration and secrets; no portable structured export/import contract was established.

### OWN-03: Data handling and telemetry

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Code and prompts stay E2EE-sealed with keys the relay never holds; credentials live in tmux env/0600 files with redacted handoffs; push goes through Apple/Google and transport over Cloudflare. Training use of Harness data, retention windows, residency, subprocessors beyond those named and telemetry controls are not disclosed.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-21. Docs state repositories and native harness credentials stay on machines you control and server state lives in a documented directory. Since 0.2.0 a daily usage report is on by default and can be turned off; crash reports stay opt-in. Usage reports carry an install id, schema version, Podium version, OS, architecture, bucketed install age and machine count, session counts per harness and whether issues were used. Crash reports carry the same install id, versions, OS and architecture, a fixed error type and up to 20 stack frames, each with a Podium source file path, line and function when there is one; error messages and frames outside Podium are dropped. Neither tier sends prompts, transcripts, your code or paths outside the Podium source tree. Model-provider handling is the harness own. Hosted analytics and website analytics are separate paths.

### OWN-04: Price and included usage

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Harness software is MIT open source and free to download with no paid plan found; agents run on your own existing subscriptions and Harness sells no tokens. No priced dimensions (currency, intervals, seats, usage, overages, taxes) are stated; hardware devices are priced separately ($129/$299) outside ADE billing.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. No Podium price is published: podium.do/pricing returns 404 and the docs list billing as not implemented. The software is open source and free to download; model and compute costs are the user's own.

### OWN-05: Scenario cost and controls

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. One, two or five developers run the free download on their own machines with their own engine subscriptions; no fixed/variable product charges exist, so scenarios carry only BYO provider/compute unknowns. No workload, concurrency, token or compute-hour assumptions are published by the vendor.
- **Podium**: No sourced public finding (not-published); support: not-published. No publishable sourced finding is available for this criterion. This does not mean the product lacks the capability.

### OWN-06: Availability and support

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. Desktop 1.4.49 and CLI 0.6.56 shipped 2026-11-13 on a near-daily cadence with public downloads and auto-update; the provider-protocol changelog runs additive/breaking revisions with deprecation service; support is GitHub issues plus site chat/community with no stated commitments; no status page or history was found; removed Swarm v2 surfaces answer empty for 30 days.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. Podium is pre-1.0. Stable v0.3.1 (2026-11-06), an update fix, follows v0.3.0 (2026-11-05), v0.2.0 (2026-10-11) and v0.1.0 (2026-08-21), with edge and dev prerelease channels. Windows and Linux desktop are Experimental, the native iPhone app is in private beta and hosted Podium is unconfirmed. Support runs through GitHub issues, Discord and private vulnerability reporting.

### OWN-07: Enterprise assurance

- **Harness by Autonomous**: No sourced public finding (not-published); support: not-published. No publishable sourced finding is available for this criterion. This does not mean the product lacks the capability.
- **Podium**: No sourced public finding (not-published); support: not-published. No publishable sourced finding is available for this criterion. This does not mean the product lacks the capability.

### OWN-08: Useful free tier and evaluation access

- **Harness by Autonomous**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-13 (9 days old); agent cross-checked 2026-11-13. The MIT desktop download is a free, ongoing tier for macOS, Windows and Linux with no paid plan, card or evaluation purchase; local harnesses need no account. Trial, credits-trial and paid-evaluation questions do not apply to a free product and are answered negatively; provider-side subscription limits still apply to agent work.
- **Podium**: Agent cross-checked (agent-cross-checked); support: partial. Checked 2026-11-06 (16 days old); agent cross-checked 2026-11-06. Podium is open-source Apache-2.0 software, and the download page calls it free and open source. Download links for macOS, Linux and Windows redirect to GitHub release assets without a login (observed by HTTP requests; not installed). No Podium charge is documented for self-hosting; agent provider usage is separate. The native iPhone app is private beta and hosted access is unconfirmed.

### OWN-09: Pricing transparency

- **Harness by Autonomous**: No sourced public finding (not-published); support: not-published. No publishable sourced finding is available for this criterion. This does not mean the product lacks the capability.
- **Podium**: No sourced public finding (not-published); support: not-published. No publishable sourced finding is available for this criterion. This does not mean the product lacks the capability.
