---
schema_version: 7
rubric_version: "0.3"
canonical_url: "https://podium.do/compare/agents/products/background-agents/"
data_url: "https://podium.do/compare/agents/products/background-agents/data.json"
generated_at: 2026-11-22
oldest_checked_at: 2026-11-08
last_checked_at: 2026-11-13
publisher: "Podium"
---

# Open-Inspect: ADE findings

> 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

40 of 51 criteria have a sourced finding; 40 were also cross-checked by a second agent. Source checks: 2026-11-08 to 2026-11-13. A missing sourced finding is not evidence that a capability is absent.

- Website: https://github.com/open-inspect/background-agents
- Canonical HTML: https://podium.do/compare/agents/products/background-agents/
- Machine-readable JSON: https://podium.do/compare/agents/products/background-agents/data.json
- Rubric: https://podium.do/compare/agents/rubric/index.md
- Comparisons with Open-Inspect on podium.do: 1 (listed in the JSON and HTML)
- Compare Open-Inspect with any other product on adecomparison.com by Podium: https://adecomparison.com/compare/?products=background-agents

## Access and agent choice

### ACC-01: Client access

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

The primary product UI is a web workspace with sessions, automations, review and settings. Changes panel has a full-width unified diff layout on phones and a small-screen session-actions menu. Full phone parity was not tested. Authorized users start sessions, queue follow-ups, stop work and review diffs/PRs through the web. The scoped user distribution is browser-only; no native desktop/phone client or session-control CLI is shipped.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- desktop_os — Desktop os [unsupported]: The complete scoped client build/distribution tree has the web workspace and no native desktop client. Sources: background-agents:gap35-clients.
- mobile_apps — Mobile apps [unsupported]: The complete scoped client build targets have no native iOS/Android app. Sources: background-agents:gap35-clients.
- browser — Browser [supported]: The primary product UI is a web workspace with sessions, automations, review and settings. Sources: background-agents:readme.
- mobile_browser — Mobile browser [partial]: Changes panel has a full-width unified diff layout on phones and a small-screen session-actions menu. Full phone parity was not tested. Sources: background-agents:review, background-agents:collaboration.
- cli — Cli [unsupported]: Deployment/admin/image scripts and MCP interfaces are supplied, without a coding-session command-line client. Sources: background-agents:gap35-clients.
- editor_extensions — Editor extensions [unknown]: Client access, editor_extensions: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- task_actions — Task actions [supported]: Authorized users start sessions, queue follow-ups, stop work and review diffs/PRs through the web. Sources: background-agents:collaboration, background-agents:review.
- notifications — Notifications [unknown]: Client access, notifications: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- cross_device_continuity — Cross device continuity [unknown]: Client access, cross_device_continuity: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-17.

Sources: background-agents:readme, background-agents:review, background-agents:collaboration, background-agents:gap35-clients

### ACC-02: Agent and model choice

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

OpenCode, Claude Agent SDK and Codex are separate selectable harnesses; each session keeps its initial harness for life. Anthropic, OpenAI, xAI and OpenCode Zen authentication/model routes are distinct from harnesses. Harnesses run inside a full development sandbox and emit a shared event stream.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- harnesses — Harnesses [supported]: OpenCode, Claude Agent and Codex are separate selectable harnesses; each session keeps its initial harness for life. The Claude Agent harness runs the bundled Claude Code binary from the Claude Agent SDK inside the sandbox. Sources: background-agents:harness, background-agents:src3-claude-binary, background-agents:src3-claude-bundled.
- model_providers — Model providers [supported]: Anthropic, OpenAI, xAI and OpenCode Zen authentication/model routes are distinct from harnesses. Sources: background-agents:harness, background-agents:accounts.
- custom_adapters — Custom adapters [unknown]: Agent and model choice, custom_adapters: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- session_integration — Session integration [supported]: Harnesses run inside a full development sandbox and emit a shared event stream. Sources: background-agents:harness.
- mixed_harness_operation — Mixed harness operation [unknown]: Agent and model choice, mixed_harness_operation: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-10.

Sources: background-agents:harness, background-agents:accounts, background-agents:src3-claude-binary, background-agents:src3-claude-bundled

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

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Claude account works on Claude Agent, ChatGPT account on supported OpenCode/Codex routes, and SuperGrok on supported OpenCode routes. Provider API keys are a separate supported authentication route. Connected account lists expose status/default and per-provider unattended authentication policy.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- agent_subscriptions — Agent subscriptions [supported]: Claude account works on Claude Agent, ChatGPT account on supported OpenCode/Codex routes, and SuperGrok on supported OpenCode routes. Sources: background-agents:harness, background-agents:accounts.
- api_accounts — Api accounts [supported]: Provider API keys are a separate supported authentication route. Sources: background-agents:accounts.
- vendor_credits — Vendor credits [unknown]: Subscriptions, API keys and vendor credits, vendor_credits: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- local_models — Local models [unknown]: Subscriptions, API keys and vendor credits, local_models: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- account_switching — Account switching [supported]: Connected account lists expose status/default and per-provider unattended authentication policy. Sources: background-agents:accounts.
- vendor_account_requirement — Vendor account requirement [unknown]: Subscriptions, API keys and vendor credits, vendor_account_requirement: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- provider_restrictions — Provider restrictions [unknown]: Subscriptions, API keys and vendor credits, provider_restrictions: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:harness, background-agents:accounts

### ACC-04: Onboarding and project prerequisites

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

The setup guide separates running the web app against an existing backend from a full self-hosted deployment, and links a container control-plane option without Cloudflare. Operators supply the GitHub App and a sandbox provider; this is an installation project rather than a managed turnkey account.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- first_task_setup — First task setup [supported]: The setup guide separates local web development against an existing backend from full-stack deployment. Sources: background-agents:setup.
- no_git — No git [unknown]: Onboarding and project prerequisites, no_git: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- local_git_without_remote — Local git without remote [unknown]: Onboarding and project prerequisites, local_git_without_remote: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- hosted_git — Hosted git [unknown]: Onboarding and project prerequisites, hosted_git: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- prerequisite_guidance — Prerequisite guidance [unknown]: Onboarding and project prerequisites, prerequisite_guidance: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- managed_cloud_setup — Managed cloud setup [partial]: Operators deploy the stack themselves, including the GitHub App and a sandbox provider, with Cloudflare or a container control plane; there is no managed turnkey account. Sources: background-agents:setup.
- byo_vps_onboarding — Byo vps onboarding [unknown]: Onboarding and project prerequisites, byo_vps_onboarding: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-17.

Sources: background-agents:setup

### ACC-05: Programmable access

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Control-plane session endpoints include budget updates and operator/service exports, with role/token authorization. The MCP server reads sessions/messages/events/diffs and optionally operates automation and skill import. PAT scope is intersected with user role and workspace policy; read-only tokens cannot write despite tool instructions.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- cli — Cli [unknown]: Programmable access, cli: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- api — Api [supported]: Control-plane session endpoints include budget updates and operator/service exports, with role/token authorization. Sources: background-agents:budget, background-agents:data.
- sdk — Sdk [unknown]: Programmable access, sdk: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- mcp_server — Mcp server [supported]: The MCP server reads sessions/messages/events/diffs and optionally operates automation and skill import. Sources: background-agents:mcp.
- webhooks — Webhooks [unknown]: Programmable access, webhooks: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- operations — Operations [unknown]: Programmable access, operations: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- authentication — Authentication [supported]: PAT scope is intersected with user role and workspace policy; read-only tokens cannot write despite tool instructions. Sources: background-agents:mcp.
- stability — Stability [unknown]: Programmable access, stability: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:budget, background-agents:data, background-agents:mcp

### ACC-06: External tools and extensions

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Managed shared/personal MCP connections are available with harness-specific loading restrictions. Claude Agent reads CLAUDE.md/.claude configuration; Codex reads AGENTS.md and refuses repository .codex/config.toml. Personal MCP proxy credentials stay outside the sandbox; loaded tool access differs by harness.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- mcp_client — Mcp client [partial]: Managed shared/personal MCP connections are available with harness-specific loading restrictions. Sources: background-agents:tools, background-agents:harness.
- plugins — Plugins [unknown]: External tools and extensions, plugins: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- custom_tools — Custom tools [unknown]: External tools and extensions, custom_tools: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- instructions_import — Instructions import [supported]: Claude Agent reads CLAUDE.md/.claude configuration; Codex reads AGENTS.md and refuses repository .codex/config.toml. Sources: background-agents:harness.
- authentication — Authentication [unknown]: External tools and extensions, authentication: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- permission_scope — Permission scope [partial]: Personal MCP proxy credentials stay outside the sandbox; loaded tool access differs by harness. Sources: background-agents:auth, background-agents:harness.
- integration_owner — Integration owner [unknown]: External tools and extensions, integration_owner: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:tools, background-agents:harness, background-agents:auth

### ACC-07: Client implementation

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

The client package is Next.js web; no native desktop shell was established.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- desktop_implementation — Desktop implementation [partial]: The client package is Next.js web; no native desktop shell was established. Sources: background-agents:readme.
- mobile_implementation — Mobile implementation [unknown]: Client implementation, mobile_implementation: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- hybrids — Hybrids [unknown]: Client implementation, hybrids: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- implementation_evidence — Implementation evidence [unknown]: Client implementation, implementation_evidence: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:readme

## Planning and context

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

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Web prompts, Slack, GitHub PRs, Linear issues and webhooks can start coding sessions. GitHub/Linear integrations remain external issue/PR records; this pass did not establish a separate native issue tracker.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- local_creation — Local creation [unknown]: Intake and external tracker relationship, local_creation: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- prompt_intake — Prompt intake [supported]: Web prompts, Slack, GitHub PRs, Linear issues and webhooks can start coding sessions. Sources: background-agents:readme.
- external_import — External import [unknown]: Intake and external tracker relationship, external_import: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- chat_git_triggers — Chat git triggers [unknown]: Intake and external tracker relationship, chat_git_triggers: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- authoritative_tracker — Authoritative tracker [partial]: GitHub/Linear integrations remain external issue/PR records; this pass did not establish a separate native issue tracker. Sources: background-agents:readme.
- stable_ids — Stable ids [unknown]: Intake and external tracker relationship, stable_ids: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- sync — Sync [unknown]: Intake and external tracker relationship, sync: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- supported_edits — Supported edits [unknown]: Intake and external tracker relationship, supported_edits: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- external_roundtrips — External roundtrips [unknown]: Intake and external tracker relationship, external_roundtrips: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:readme

### PLAN-02: Durable task plans

Verification status: 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.

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

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Sessions retain child/grandchild relationships with a bounded depth and separate sandbox/branch per child.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- epics — Epics [unknown]: Hierarchy, dependencies and evolving scope, epics: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- hierarchy — Hierarchy [partial]: Sessions retain child/grandchild relationships with a bounded depth and separate sandbox/branch per child. Sources: background-agents:children.
- blockers — Blockers [unknown]: Hierarchy, dependencies and evolving scope, blockers: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- readiness — Readiness [unknown]: Hierarchy, dependencies and evolving scope, readiness: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- enforcement — Enforcement [unknown]: Hierarchy, dependencies and evolving scope, enforcement: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- state_propagation — State propagation [unknown]: Hierarchy, dependencies and evolving scope, state_propagation: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- task_to_epic — Task to epic [unknown]: Hierarchy, dependencies and evolving scope, task_to_epic: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- identity_retention — Identity retention [unknown]: Hierarchy, dependencies and evolving scope, identity_retention: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- workflow_constraints — Workflow constraints [unknown]: Hierarchy, dependencies and evolving scope, workflow_constraints: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:children

### PLAN-04: Shared context and memory

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Harness-native repository instruction files and pinned managed skills provide session context. Scoped persistent memory is configured separately from session transcripts. No living-product-spec approval system is established. Personal, repository and environment memories have separate permissions and session-loading rules. Humans edit/approve scoped memories; repository/environment agent proposals need approval before later sessions use them. Existing session context is not retroactively erased.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- instructions — Instructions [supported]: Harness-native repository instruction files and pinned managed skills provide session context. Sources: background-agents:harness, background-agents:children.
- skills — Skills [unknown]: Shared context and memory, skills: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- project_docs — Project docs [unknown]: Shared context and memory, project_docs: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- memory — Memory [partial]: Scoped persistent memory is configured separately from session transcripts. No living-product-spec approval system is established. Sources: background-agents:memory.
- scope — Scope [supported]: Personal, repository and environment memories have separate permissions and session-loading rules. Sources: background-agents:memory.
- provenance — Provenance [unknown]: Shared context and memory, provenance: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- correction — Correction [supported]: Humans edit/approve scoped memories; repository/environment agent proposals need approval before later sessions use them. Existing session context is not retroactively erased. Sources: background-agents:memory.
- deletion — Deletion [unknown]: Shared context and memory, deletion: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- freshness — Freshness [unknown]: Shared context and memory, freshness: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- takeover_continuity — Takeover continuity [unknown]: Shared context and memory, takeover_continuity: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:harness, background-agents:children, background-agents:memory

### PLAN-05: Architecture and API awareness

Verification status: 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.

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

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Native inbox, searches and child-task cards are session records. The scoped data/API model has no independent product-owned issue/task entity; Linear/GitHub are external integrations.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Source inspection covers the scoped release only. Unanswered atomic questions remain unknown; no live scenario was run.

- create — Create [unknown]: Native issue management and idea backlog, create: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- edit — Edit [unknown]: Native issue management and idea backlog, edit: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- discuss — Discuss [unknown]: Native issue management and idea backlog, discuss: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- prioritize — Prioritize [unknown]: Native issue management and idea backlog, prioritize: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- label — Label [unknown]: Native issue management and idea backlog, label: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- assign — Assign [unknown]: Native issue management and idea backlog, assign: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- search — Search [unknown]: Native issue management and idea backlog, search: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- filter — Filter [unknown]: Native issue management and idea backlog, filter: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- archive — Archive [unknown]: Native issue management and idea backlog, archive: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- reopen — Reopen [unknown]: Native issue management and idea backlog, reopen: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- project_epic_views — Project epic views [unknown]: Native issue management and idea backlog, project_epic_views: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- idea_parking — Idea parking [unknown]: Native issue management and idea backlog, idea_parking: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- explicit_dispatch — Explicit dispatch [unknown]: Native issue management and idea backlog, explicit_dispatch: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- task_session_separation — Task session separation [unsupported]: Canonical schemas/routes model sessions and external tracker references, without independent issue/task CRUD. Sources: background-agents:gap35-native-issues.
- metadata — Metadata [unknown]: Native issue management and idea backlog, metadata: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- durable_states — Durable states [unknown]: Native issue management and idea backlog, durable_states: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- external_tracker_requirement — External tracker requirement [unknown]: Native issue management and idea backlog, external_tracker_requirement: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-13 (9 days old on 2026-11-22); agent cross-checked 2026-11-17.

Sources: background-agents:gap35-native-issues

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

Verification status: 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.

### PLAN-08: Living product specifications

Verification status: 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.

## Coordination and automation

### COORD-01: Parallel work

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Parent continues while independent child sandboxes run simultaneously. Default limits allow five concurrent and fifteen total children per parent.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- concurrent_sessions — Concurrent sessions [supported]: Parent continues while independent child sandboxes run simultaneously. Sources: background-agents:children.
- concurrent_tasks — Concurrent tasks [unknown]: Parallel work, concurrent_tasks: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- quotas — Quotas [supported]: Default limits allow five concurrent and fifteen total children per parent. Sources: background-agents:children.
- observed_concurrency — Observed concurrency [unknown]: Parallel work, observed_concurrency: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- status_visibility — Status visibility [unknown]: Parallel work, status_visibility: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:children

### COORD-02: Ownership and dispatch

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Automation admission rechecks executor/target permission; denied scheduled authorization pauses the automation.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- human_assignment — Human assignment [unknown]: Ownership and dispatch, human_assignment: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- agent_assignment — Agent assignment [unknown]: Ownership and dispatch, agent_assignment: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- claims — Claims [unknown]: Ownership and dispatch, claims: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- queues — Queues [unknown]: Ownership and dispatch, queues: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- routing — Routing [supported]: Automation admission rechecks executor/target permission; denied scheduled authorization pauses the automation. Sources: background-agents:automation.
- duplicate_prevention — Duplicate prevention [unknown]: Ownership and dispatch, duplicate_prevention: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- ownership_enforcement — Ownership enforcement [unknown]: Ownership and dispatch, ownership_enforcement: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:automation

### COORD-03: Inspectable delegation and communication

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Child can override model/reasoning but inherits its parents harness and authentication. Product-owned spawn-child/send-child-prompt/get-child-status/cancel-child tools manage independent sessions. UI nests children, shows live status/PR state and links parent and child pages. get-child-status can retrieve final responses and paginated event history.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- roles — Roles [unknown]: Inspectable delegation and communication, roles: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- model_harness_effort — Model harness effort [supported]: Child can override model/reasoning but inherits its parents harness and authentication. Sources: background-agents:children.
- messages — Messages [unknown]: Inspectable delegation and communication, messages: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- results_return — Results return [unknown]: Inspectable delegation and communication, results_return: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- cross_harness — Cross harness [unknown]: Inspectable delegation and communication, cross_harness: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- orchestration_owner — Orchestration owner [supported]: Product-owned spawn-child/send-child-prompt/get-child-status/cancel-child tools manage independent sessions. Sources: background-agents:children.
- activation — Activation [unknown]: Inspectable delegation and communication, activation: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- authentication — Authentication [unknown]: Inspectable delegation and communication, authentication: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- live_progress — Live progress [supported]: UI nests children, shows live status/PR state and links parent and child pages. Sources: background-agents:children.
- retained_results — Retained results [supported]: get-child-status can retrieve final responses and paginated event history. Sources: background-agents:children.
- followup — Followup [unknown]: Inspectable delegation and communication, followup: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- invocation — Invocation [unknown]: Inspectable delegation and communication, invocation: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:children

### COORD-04: Unattended triggers

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Cron automation starts ordinary sessions without a human prompting each run. Inbound webhook, Sentry, Slack and configured GitHub events trigger automation; Linear automation events are still planned. Three consecutive run failures pause an automation; scheduled authorization denial pauses immediately.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- schedules — Schedules [supported]: Cron automation starts ordinary sessions without a human prompting each run. Sources: background-agents:automation.
- events — Events [partial]: Inbound webhook, Sentry, Slack and configured GitHub events trigger automation; Linear automation events are still planned. Sources: background-agents:automation.
- recurring_tasks — Recurring tasks [unknown]: Unattended triggers, recurring_tasks: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- completion_conditions — Completion conditions [unknown]: Unattended triggers, completion_conditions: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- scheduler_location — Scheduler location [unknown]: Unattended triggers, scheduler_location: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- overlaps — Overlaps [unknown]: Unattended triggers, overlaps: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- retries — Retries [supported]: Three consecutive run failures pause an automation; scheduled authorization denial pauses immediately. Sources: background-agents:automation.
- client_independence — Client independence [unknown]: Unattended triggers, client_independence: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:automation

### COORD-05: Intervention and attention

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Authorized collaborators can stop the current session turn. Follow-ups from any authorized human queue in order behind the current turn.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- pause — Pause [unknown]: Intervention and attention, pause: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- stop — Stop [supported]: Authorized collaborators can stop the current session turn. Sources: background-agents:collaboration.
- resume — Resume [unknown]: Intervention and attention, resume: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- steer — Steer [supported]: Follow-ups from any authorized human queue in order behind the current turn. Sources: background-agents:collaboration.
- notifications — Notifications [unknown]: Intervention and attention, notifications: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- pending_decisions — Pending decisions [unknown]: Intervention and attention, pending_decisions: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- escalation — Escalation [unknown]: Intervention and attention, escalation: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- visible_errors — Visible errors [unknown]: Intervention and attention, visible_errors: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- chat_terminal_parity — Chat terminal parity [unknown]: Intervention and attention, chat_terminal_parity: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- retained_history — Retained history [unknown]: Intervention and attention, retained_history: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:collaboration

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

Verification status: 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.

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

Verification status: 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.

## Execution and reliability

### RUN-01: Laptop-independent execution

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Operators self-host the control plane and configure their own remote sandbox-provider accounts for agent execution. The supported backend list has no local-laptop agent runtime or Open-Inspect-vendor compute service.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- local — Local [unsupported]: The complete selectable agent backend list is remote provider sandboxes; local control-plane hosting is not local agent execution. Sources: background-agents:gap35-execution.
- byo_remote — Byo remote [supported]: Operators use their own sandbox-provider accounts and infrastructure configuration. Sources: background-agents:gap35-execution.
- self_hosted — Self hosted [supported]: Operators deploy the full self-hosted stack, with the control plane on Cloudflare Workers or in a container. Sources: background-agents:setup.
- managed_cloud — Managed cloud [unsupported]: Open-Inspect supplies self-hosted code, not a vendor-operated compute service; sandbox vendors are the operator chosen infrastructure providers. Sources: background-agents:gap35-execution.
- hybrid — Hybrid [unknown]: Laptop-independent execution, hybrid: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- provisioning — Provisioning [unknown]: Laptop-independent execution, provisioning: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- payer — Payer [unknown]: Laptop-independent execution, payer: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- resources — Resources [unknown]: Laptop-independent execution, resources: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- region — Region [unknown]: Laptop-independent execution, region: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- expiry — Expiry [unknown]: Laptop-independent execution, expiry: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- client_disconnect — Client disconnect [supported]: The browser is a control client; background sandbox work persists and a colleague can continue the same session after the original user leaves. Sources: background-agents:readme, background-agents:collaboration.
- other_device_continuation — Other device continuation [unknown]: Laptop-independent execution, other_device_continuation: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-17.

Sources: background-agents:setup, background-agents:readme, background-agents:sandbox, background-agents:collaboration, background-agents:gap35-execution

### RUN-02: Workspace and process isolation

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Each child gets its own branch and sandbox rather than the parents working branch. Default development runtime uses a container; modal-vm and Boat are separately documented VM-backed options. modal-vm and Boat provide explicit VM-backed execution options. Separate sandbox checkouts/branches do not use Git worktree provisioning in this revision.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- checkout — Checkout [unknown]: Workspace and process isolation, checkout: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- branch — Branch [supported]: Each child gets its own branch and sandbox rather than the parents working branch. Sources: background-agents:children.
- worktree — Worktree [unsupported]: The complete sandbox repository boot/sync path clones checkouts; no git worktree provisioning path is used. Sources: background-agents:gap35-checkouts.
- container — Container [supported]: Default development runtime uses a container; modal-vm and Boat are separately documented VM-backed options. Sources: background-agents:runtime.
- vm — Vm [supported]: modal-vm and Boat provide explicit VM-backed execution options. Sources: background-agents:sandbox.
- os_sandbox — Os sandbox [unknown]: Workspace and process isolation, os_sandbox: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- filesystem — Filesystem [unknown]: Workspace and process isolation, filesystem: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- process — Process [unknown]: Workspace and process isolation, process: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- network — Network [unknown]: Workspace and process isolation, network: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- secrets — Secrets [unknown]: Workspace and process isolation, secrets: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- code_separation — Code separation [unknown]: Workspace and process isolation, code_separation: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- security_boundary — Security boundary [unknown]: Workspace and process isolation, security_boundary: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-17.

Sources: background-agents:children, background-agents:runtime, background-agents:sandbox, background-agents:gap35-checkouts

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

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Provider-specific expiry can snapshot/restore or stop/pause/archive and later resume retained state. Each harness has its own session/thread resume identifier; sandbox replacement is distinct from live process reconnection.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- client_disconnect — Client disconnect [unknown]: Recovery, expiry and continuity, client_disconnect: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- worker_crash — Worker crash [unknown]: Recovery, expiry and continuity, worker_crash: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- service_restart — Service restart [unknown]: Recovery, expiry and continuity, service_restart: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- expiry — Expiry [partial]: Provider-specific expiry can snapshot/restore or stop/pause/archive and later resume retained state. Sources: background-agents:sandbox.
- transcript_reload — Transcript reload [unknown]: Recovery, expiry and continuity, transcript_reload: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- process_reconnect — Process reconnect [unknown]: Recovery, expiry and continuity, process_reconnect: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- task_resume — Task resume [unknown]: Recovery, expiry and continuity, task_resume: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- state_reconstruction — State reconstruction [unknown]: Recovery, expiry and continuity, state_reconstruction: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- context_loss — Context loss [partial]: Each harness has its own session/thread resume identifier; sandbox replacement is distinct from live process reconnection. Sources: background-agents:harness, background-agents:sandbox.
- duplicate_actions — Duplicate actions [unknown]: Recovery, expiry and continuity, duplicate_actions: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- manual_repair — Manual repair [unknown]: Recovery, expiry and continuity, manual_repair: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- multi_day_continuity — Multi day continuity [unknown]: Recovery, expiry and continuity, multi_day_continuity: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:sandbox, background-agents:harness

### RUN-04: Development environment reliability

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Prebuilt images and saved-state models differ by provider; Modal supports per-turn snapshots while several backends snapshot at shutdown. Session GitHub tokens are scoped to persisted repositories/team grants, while cached credentials can remain in snapshots.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- dependencies — Dependencies [unknown]: Development environment reliability, dependencies: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- caches — Caches [unknown]: Development environment reliability, caches: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- snapshots — Snapshots [supported]: Prebuilt images and saved-state models differ by provider; Modal supports per-turn snapshots while several backends snapshot at shutdown. Sources: background-agents:sandbox.
- services — Services [unknown]: Development environment reliability, services: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- ports — Ports [unknown]: Development environment reliability, ports: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- secrets — Secrets [partial]: Session GitHub tokens are scoped to persisted repositories/team grants, while cached credentials can remain in snapshots. Sources: background-agents:auth.
- duplicate_installs — Duplicate installs [unknown]: Development environment reliability, duplicate_installs: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- collisions — Collisions [unknown]: Development environment reliability, collisions: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- remote_setup — Remote setup [unknown]: Development environment reliability, remote_setup: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- automatic_repair — Automatic repair [unknown]: Development environment reliability, automatic_repair: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:sandbox, background-agents:auth

### RUN-05: Limits and failure handling

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

A reported USD session limit stops active work and blocks queued/new prompts until the owner raises/removes it. Limits document a four-minute connect watchdog, 30-minute default boot budget and 90-second heartbeat timeout with a ready-runtime reconnect grace. Limits apply at reporting boundaries and cannot bound unreported model usage or guarantee exact-dollar stop.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- time_limits — Time limits [unknown]: Limits and failure handling, time_limits: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- turn_limits — Turn limits [unknown]: Limits and failure handling, turn_limits: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- token_limits — Token limits [unknown]: Limits and failure handling, token_limits: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- spend_limits — Spend limits [supported]: A reported USD session limit stops active work and blocks queued/new prompts until the owner raises/removes it. Sources: background-agents:budget.
- timeouts — Timeouts [supported]: Limits document a four-minute connect watchdog, 30-minute default boot budget and 90-second heartbeat timeout with a ready-runtime reconnect grace. Sources: background-agents:limits.
- retries — Retries [unknown]: Limits and failure handling, retries: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- provider_failures — Provider failures [unknown]: Limits and failure handling, provider_failures: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- enforcement — Enforcement [partial]: Limits apply at reporting boundaries and cannot bound unreported model usage or guarantee exact-dollar stop. Sources: background-agents:budget.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:budget, background-agents:limits

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

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Sessions share token/tool/step/warning events across harnesses; child history can be retrieved programmatically.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- logs — Logs [supported]: Sessions share token/tool/step/warning events across harnesses; child history can be retrieved programmatically. Sources: background-agents:harness, background-agents:children.
- traces — Traces [unknown]: Truthful status and operational statistics, traces: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- failure_reasons — Failure reasons [unknown]: Truthful status and operational statistics, failure_reasons: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- usage_export — Usage export [unknown]: Truthful status and operational statistics, usage_export: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- task_stats — Task stats [unknown]: Truthful status and operational statistics, task_stats: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- agent_stats — Agent stats [unknown]: Truthful status and operational statistics, agent_stats: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- project_stats — Project stats [unknown]: Truthful status and operational statistics, project_stats: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- time_stats — Time stats [unknown]: Truthful status and operational statistics, time_stats: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- failure_status — Failure status [unknown]: Truthful status and operational statistics, failure_status: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- approval_status — Approval status [unknown]: Truthful status and operational statistics, approval_status: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- reconnect_consistency — Reconnect consistency [unknown]: Truthful status and operational statistics, reconnect_consistency: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- performance_method — Performance method [unknown]: Truthful status and operational statistics, performance_method: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- responsiveness — Responsiveness [unknown]: Truthful status and operational statistics, responsiveness: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:harness, background-agents:children

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

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Budget sidebar displays reported session cost and configured limit. Claude SDK/Codex subscription figures are informational estimates. Missing Codex prices may report zero with a warning, not a measured zero bill.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- task_cost — Task cost [unknown]: Per-task API and compute cost, task_cost: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- session_cost — Session cost [supported]: Budget sidebar displays reported session cost and configured limit. Sources: background-agents:budget.
- child_cost — Child cost [unknown]: Per-task API and compute cost, child_cost: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- epic_cost — Epic cost [unknown]: Per-task API and compute cost, epic_cost: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- harness_aggregation — Harness aggregation [unknown]: Per-task API and compute cost, harness_aggregation: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- billed_vs_estimated — Billed vs estimated [partial]: Claude SDK/Codex subscription figures are informational estimates. Missing Codex prices may report zero with a warning, not a measured zero bill. Sources: background-agents:budget.
- tokens — Tokens [unknown]: Per-task API and compute cost, tokens: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- subscription_allowance — Subscription allowance [unknown]: Per-task API and compute cost, subscription_allowance: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- provider_reconciliation — Provider reconciliation [unknown]: Per-task API and compute cost, provider_reconciliation: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- compute_cost — Compute cost [unknown]: Per-task API and compute cost, compute_cost: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- refresh_delay — Refresh delay [unknown]: Per-task API and compute cost, refresh_delay: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- alerts — Alerts [unknown]: Per-task API and compute cost, alerts: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- export — Export [unknown]: Per-task API and compute cost, export: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- drilldown — Drilldown [unknown]: Per-task API and compute cost, drilldown: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:budget

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

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Child spawning fails when the configured concurrent limit is reached; it is not described as silently queued beyond that limit. Provider-specific CPU/memory overrides are honored on some backends and dropped on Daytona/E2B.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- concurrency — Concurrency [partial]: Child spawning fails when the configured concurrent limit is reached; it is not described as silently queued beyond that limit. Sources: background-agents:children.
- queues — Queues [unknown]: Compute-aware scheduling and verification, queues: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- cpu_ram — Cpu ram [partial]: Provider-specific CPU/memory overrides are honored on some backends and dropped on Daytona/E2B. Sources: background-agents:sandbox.
- backpressure — Backpressure [unknown]: Compute-aware scheduling and verification, backpressure: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- host_awareness — Host awareness [unknown]: Compute-aware scheduling and verification, host_awareness: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- check_reuse — Check reuse [unknown]: Compute-aware scheduling and verification, check_reuse: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- check_deduplication — Check deduplication [unknown]: Compute-aware scheduling and verification, check_deduplication: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- bounded_tests — Bounded tests [unknown]: Compute-aware scheduling and verification, bounded_tests: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- cancellation — Cancellation [unknown]: Compute-aware scheduling and verification, cancellation: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- fairness — Fairness [unknown]: Compute-aware scheduling and verification, fairness: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- recovery — Recovery [unknown]: Compute-aware scheduling and verification, recovery: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- shared_limits — Shared limits [unknown]: Compute-aware scheduling and verification, shared_limits: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:children, background-agents:sandbox

## Human collaboration and control

### TEAM-01: Live human collaboration

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Participants and connected-presence dots update for viewers in the same session. Multiple authorized people can send follow-ups to the same session; ordering is a shared queue. Prompt author identity flows into Git commit attribution and connected-user PR identity.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- visibility — Visibility [unknown]: Live human collaboration, visibility: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- discussion — Discussion [unknown]: Live human collaboration, discussion: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- comments — Comments [unknown]: Live human collaboration, comments: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- presence — Presence [supported]: Participants and connected-presence dots update for viewers in the same session. Sources: background-agents:collaboration.
- coprompting — Coprompting [supported]: Multiple authorized people can send follow-ups to the same session; ordering is a shared queue. Sources: background-agents:collaboration.
- ordering — Ordering [unknown]: Live human collaboration, ordering: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- conflicts — Conflicts [unknown]: Live human collaboration, conflicts: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- attribution — Attribution [supported]: Prompt author identity flows into Git commit attribution and connected-user PR identity. Sources: background-agents:collaboration.
- permissions — Permissions [unknown]: Live human collaboration, permissions: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- plan_restrictions — Plan restrictions [unknown]: Live human collaboration, plan_restrictions: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- owner_online — Owner online [unknown]: Live human collaboration, owner_online: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:collaboration

### TEAM-02: Roles and access boundaries

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Owner, Administrator, Member and Viewer have documented workspace permissions. Team membership and resource-specific checks supplement workspace role; team read isolation requires enforcement on rather than default shadow. Member suspension and team removal affect future authorized operations but do not instantly revoke cached issued tokens or stop already executing sandboxes. Built-in sign-in uses Google/GitHub OAuth without enterprise SAML/OIDC SSO configuration in this scoped release.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- organization_roles — Organization roles [supported]: Owner, Administrator, Member and Viewer have documented workspace permissions. Sources: background-agents:roles.
- project_roles — Project roles [partial]: Team membership and resource-specific checks supplement workspace role; team read isolation requires enforcement on rather than default shadow. Sources: background-agents:roles, background-agents:collaboration.
- task_roles — Task roles [unknown]: Roles and access boundaries, task_roles: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- membership — Membership [unknown]: Roles and access boundaries, membership: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- sso — Sso [unsupported]: Complete sign-in config supports Google/GitHub social OAuth, with no enterprise SSO provider/plugin. Sources: background-agents:gap35-auth.
- provisioning — Provisioning [unknown]: Roles and access boundaries, provisioning: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- unauthorized_access — Unauthorized access [unknown]: Roles and access boundaries, unauthorized_access: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- revocation — Revocation [partial]: Member suspension and team removal affect future authorized operations but do not instantly revoke cached issued tokens or stop already executing sandboxes. Sources: background-agents:auth.
- plan_scope — Plan scope [unknown]: Roles and access boundaries, plan_scope: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-17.

Sources: background-agents:roles, background-agents:collaboration, background-agents:auth, background-agents:gap35-auth

### TEAM-03: Approvals and temporary access

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Credential brokering caches data on disk and snapshots can retain it; GitLab uses a deployment PAT.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- command — Command [unknown]: Approvals and temporary access, command: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- tool — Tool [unknown]: Approvals and temporary access, tool: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- network — Network [unknown]: Approvals and temporary access, network: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- secret — Secret [partial]: Credential brokering caches data on disk and snapshots can retain it; GitLab uses a deployment PAT. Sources: background-agents:auth.
- delivery — Delivery [unknown]: Approvals and temporary access, delivery: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- approver_identity — Approver identity [unknown]: Approvals and temporary access, approver_identity: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- lifetime — Lifetime [unknown]: Approvals and temporary access, lifetime: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- allow — Allow [unknown]: Approvals and temporary access, allow: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- deny — Deny [unknown]: Approvals and temporary access, deny: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- revoke — Revoke [unknown]: Approvals and temporary access, revoke: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- tunnels — Tunnels [unknown]: Approvals and temporary access, tunnels: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- defaults — Defaults [unknown]: Approvals and temporary access, defaults: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- safer_modes — Safer modes [unknown]: Approvals and temporary access, safer_modes: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- enforcement — Enforcement [unknown]: Approvals and temporary access, enforcement: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:auth

### TEAM-04: Decision and activity history

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

The audit UI records operation/authorization outcomes with actor/resource snapshots; Allowed is not evidence the operation succeeded. Operators can export paginated NDJSON traces with session_error/cursor/error records; not a consistent backup.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- actions — Actions [partial]: The audit UI records operation/authorization outcomes with actor/resource snapshots; Allowed is not evidence the operation succeeded. Sources: background-agents:audit.
- approvals — Approvals [unknown]: Decision and activity history, approvals: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- edits — Edits [unknown]: Decision and activity history, edits: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- discussion — Discussion [unknown]: Decision and activity history, discussion: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- decisions — Decisions [unknown]: Decision and activity history, decisions: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- cross_user_persistence — Cross user persistence [unknown]: Decision and activity history, cross_user_persistence: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- cross_session_persistence — Cross session persistence [unknown]: Decision and activity history, cross_session_persistence: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- export — Export [partial]: Operators can export paginated NDJSON traces with session_error/cursor/error records; not a consistent backup. Sources: background-agents:data.
- mutability — Mutability [unknown]: Decision and activity history, mutability: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- retention — Retention [unknown]: Decision and activity history, retention: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- integrity — Integrity [unknown]: Decision and activity history, integrity: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:audit, background-agents:data

### TEAM-05: Review experience and guidance

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Changes panel has path filters, unified/split views and Markdown preview with session-start/head context. Phones use full-width Unified diffs; Split is available only at sufficient width.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- pending_reviews — Pending reviews [unknown]: Review experience and guidance, pending_reviews: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- change_understanding — Change understanding [supported]: Changes panel has path filters, unified/split views and Markdown preview with session-start/head context. Sources: background-agents:review.
- evidence_navigation — Evidence navigation [unknown]: Review experience and guidance, evidence_navigation: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- comments — Comments [unknown]: Review experience and guidance, comments: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- request_changes — Request changes [unknown]: Review experience and guidance, request_changes: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- shipping_guidance — Shipping guidance [unknown]: Review experience and guidance, shipping_guidance: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- effort — Effort [unknown]: Review experience and guidance, effort: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- missed_evidence — Missed evidence [unknown]: Review experience and guidance, missed_evidence: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- keyboard — Keyboard [unknown]: Review experience and guidance, keyboard: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- narrow_screen — Narrow screen [supported]: Phones use full-width Unified diffs; Split is available only at sufficient width. Sources: background-agents:review.
- context_switching — Context switching [unknown]: Review experience and guidance, context_switching: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- roles — Roles [unknown]: Review experience and guidance, roles: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- retained_context — Retained context [unknown]: Review experience and guidance, retained_context: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:review

### TEAM-06: Colleague takeover and continuity

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

An authorized colleague can open history, queued prompts and current sandbox state and continue the same branch. This is documented, not observed with two accounts.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- ongoing_takeover — Ongoing takeover [unknown]: Colleague takeover and continuity, ongoing_takeover: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- same_task — Same task [partial]: An authorized colleague can open history, queued prompts and current sandbox state and continue the same branch. This is documented, not observed with two accounts. Sources: background-agents:collaboration.
- session_history — Session history [unknown]: Colleague takeover and continuity, session_history: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- decisions — Decisions [unknown]: Colleague takeover and continuity, decisions: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- artifacts — Artifacts [unknown]: Colleague takeover and continuity, artifacts: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- permissions — Permissions [unknown]: Colleague takeover and continuity, permissions: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- runtime_context — Runtime context [unknown]: Colleague takeover and continuity, runtime_context: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- owner_disconnect — Owner disconnect [unknown]: Colleague takeover and continuity, owner_disconnect: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- ownership — Ownership [unknown]: Colleague takeover and continuity, ownership: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- attribution — Attribution [unknown]: Colleague takeover and continuity, attribution: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- context_loss — Context loss [unknown]: Colleague takeover and continuity, context_loss: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- harness_transfer — Harness transfer [unknown]: Colleague takeover and continuity, harness_transfer: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:collaboration

## Verification and delivery

### SHIP-01: Retained artifacts and outputs

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Changes includes committed/uncommitted tracked and nonignored untracked files since session start, with named omission/error states. Timeline retains screenshot cards and recordings alongside tool output. Parent can retrieve childs final response and event history.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- diffs — Diffs [supported]: Changes includes committed/uncommitted tracked and nonignored untracked files since session start, with named omission/error states. Sources: background-agents:review.
- documents — Documents [unknown]: Retained artifacts and outputs, documents: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- screenshots — Screenshots [supported]: Timeline retains screenshot cards and recordings alongside tool output. Sources: background-agents:review.
- recordings — Recordings [unknown]: Retained artifacts and outputs, recordings: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- previews — Previews [unknown]: Retained artifacts and outputs, previews: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- test_outputs — Test outputs [unknown]: Retained artifacts and outputs, test_outputs: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- reviewer_outputs — Reviewer outputs [unknown]: Retained artifacts and outputs, reviewer_outputs: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- gallery — Gallery [unknown]: Retained artifacts and outputs, gallery: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- task_links — Task links [unknown]: Retained artifacts and outputs, task_links: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- revision_links — Revision links [unknown]: Retained artifacts and outputs, revision_links: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- persistence — Persistence [unknown]: Retained artifacts and outputs, persistence: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- child_findings — Child findings [supported]: Parent can retrieve childs final response and event history. Sources: background-agents:children.
- prompt_leakage — Prompt leakage [unknown]: Retained artifacts and outputs, prompt_leakage: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:review, background-agents:children

### SHIP-02: Verification and spec gates

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

GitHub PRs and checks are review evidence; a failed-test or requirement-conflict shipping gate was not exercised.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- tests — Tests [unknown]: Verification and spec gates, tests: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- lint — Lint [unknown]: Verification and spec gates, lint: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- types — Types [unknown]: Verification and spec gates, types: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- ci — Ci [partial]: GitHub PRs and checks are review evidence; a failed-test or requirement-conflict shipping gate was not exercised. Sources: background-agents:review.
- independent_review — Independent review [unknown]: Verification and spec gates, independent_review: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- spec_checks — Spec checks [unknown]: Verification and spec gates, spec_checks: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- required_checks — Required checks [unknown]: Verification and spec gates, required_checks: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- failure_blocking — Failure blocking [unknown]: Verification and spec gates, failure_blocking: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- requirement_conflicts — Requirement conflicts [unknown]: Verification and spec gates, requirement_conflicts: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- configured_policy — Configured policy [unknown]: Verification and spec gates, configured_policy: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- enforced_policy — Enforced policy [unknown]: Verification and spec gates, enforced_policy: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:review

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

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Each child branches from the parents base branch in its own sandbox. Sessions and children can create PRs; human prompt attribution differs from automation App-authored PRs.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- branches — Branches [supported]: Each child branches from the parents base branch in its own sandbox. Sources: background-agents:children.
- worktrees — Worktrees [unknown]: Git and pull request workflow, worktrees: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- commits — Commits [unknown]: Git and pull request workflow, commits: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- pull_requests — Pull requests [supported]: Sessions and children can create PRs; human prompt attribution differs from automation App-authored PRs. Sources: background-agents:readme, background-agents:automation, background-agents:collaboration.
- comments — Comments [unknown]: Git and pull request workflow, comments: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- ci_response — Ci response [unknown]: Git and pull request workflow, ci_response: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- rebase — Rebase [unknown]: Git and pull request workflow, rebase: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- conflicts — Conflicts [unknown]: Git and pull request workflow, conflicts: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- product_state — Product state [unknown]: Git and pull request workflow, product_state: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:children, background-agents:readme, background-agents:automation, background-agents:collaboration

### SHIP-04: Merge and deployment authority

Verification status: 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.

### SHIP-05: Completion evidence and rollback

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Child UI and session review expose linked pull request state and navigation.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- commit_links — Commit links [unknown]: Completion evidence and rollback, commit_links: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- pr_links — Pr links [supported]: Child UI and session review expose linked pull request state and navigation. Sources: background-agents:children, background-agents:review.
- build_links — Build links [unknown]: Completion evidence and rollback, build_links: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- deployment_links — Deployment links [unknown]: Completion evidence and rollback, deployment_links: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- delivery_status — Delivery status [unknown]: Completion evidence and rollback, delivery_status: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- retained_evidence — Retained evidence [unknown]: Completion evidence and rollback, retained_evidence: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- rollback — Rollback [unknown]: Completion evidence and rollback, rollback: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- revert — Revert [unknown]: Completion evidence and rollback, revert: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- stale_approval_invalidation — Stale approval invalidation [unknown]: Completion evidence and rollback, stale_approval_invalidation: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:children, background-agents:review

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

Verification status: 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.

## Ownership, privacy and economics

### OWN-01: Source and license

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

The web/control-plane application source uses MIT. The same MIT repository grant covers backend code.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- client_license — Client license [supported]: The web/control-plane application source uses MIT. Sources: background-agents:license.
- server_license — Server license [supported]: The same MIT repository grant covers backend code. Sources: background-agents:license.
- hosted_extensions — Hosted extensions [unknown]: Source and license, hosted_extensions: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- source_visibility — Source visibility [unknown]: Source and license, source_visibility: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- source_revision — Source revision [unknown]: Source and license, source_revision: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- hosted_parity — Hosted parity [unknown]: Source and license, hosted_parity: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:license

### OWN-02: Hosting and exit

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Full self-hosted deployment is documented, with the control plane on Cloudflare Workers by default or in a container. It still depends on GitHub App access, a chosen sandbox provider and model accounts. Session NDJSON export is operator/service-facing and best-effort rather than an app export/restore flow.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- self_hosting — Self hosting [supported]: Operators can deploy the full stack themselves, with the control plane on Cloudflare Workers or in a container. Sources: background-agents:setup.
- vendor_dependencies — Vendor dependencies [supported]: Self-deployment depends on GitHub App access, a chosen sandbox provider, model accounts, and Cloudflare unless the container control plane is used. Sources: background-agents:setup, background-agents:readme.
- export — Export [partial]: Session NDJSON export is operator/service-facing and best-effort rather than an app export/restore flow. Sources: background-agents:data.
- import — Import [unknown]: Hosting and exit, import: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- formats — Formats [unknown]: Hosting and exit, formats: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- cancellation — Cancellation [unknown]: Hosting and exit, cancellation: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- account_independence — Account independence [unknown]: Hosting and exit, account_independence: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- history_portability — History portability [unknown]: Hosting and exit, history_portability: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-10.

Sources: background-agents:setup, background-agents:readme, background-agents:data

### OWN-03: Data handling and telemetry

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Snapshots can retain on-disk cached credentials; removing a grant does not instantly revoke issued tokens. Web exposes archive/unarchive rather than deletion; control-plane delete exists but has no app control.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement. Source-depth answers added 2026-11-16 from release main@531aedf (531aedffb672); older citations identify their own revisions.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- code_data — Code data [unknown]: Data handling and telemetry, code_data: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- prompts — Prompts [unknown]: Data handling and telemetry, prompts: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- credentials — Credentials [partial]: Snapshots can retain on-disk cached credentials; removing a grant does not instantly revoke issued tokens. Sources: background-agents:auth.
- logs — Logs [unknown]: Data handling and telemetry, logs: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- artifacts — Artifacts [unknown]: Data handling and telemetry, artifacts: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- retention — Retention [unknown]: Data handling and telemetry, retention: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- deletion — Deletion [partial]: Web exposes archive/unarchive rather than deletion; control-plane delete exists but has no app control. Sources: background-agents:data.
- training — Training [unknown]: Data handling and telemetry, training: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- residency — Residency [unknown]: Data handling and telemetry, residency: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- subprocessors — Subprocessors [unknown]: Data handling and telemetry, subprocessors: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- telemetry_disclosure — Telemetry disclosure [unsupported]: No telemetry is sent to the maintainers, so none is disclosed; the in-product analytics page shows the operator's own usage. Sources: background-agents:v531-n-telemetry.
- telemetry_defaults — Telemetry defaults [supported]: No vendor telemetry code exists, so nothing is sent by default. Sources: background-agents:v531-n-telemetry.
- telemetry_controls — Telemetry controls [unsupported]: No telemetry opt-out control exists because no vendor telemetry exists. Sources: background-agents:v531-n-telemetry.
- consent_prompt — Consent prompt [unsupported]: No telemetry means no consent prompt. Sources: background-agents:v531-n-telemetry.
- provider_separation — Provider separation [unknown]: Data handling and telemetry, provider_separation: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-17.

Sources: background-agents:auth, background-agents:data, background-agents:v531-n-telemetry

### OWN-04: Price and included usage

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Open-Inspect is free, self-hosted MIT software with no paid plan found; model usage is billed through the user's own API key or subscriptions, and sandbox/edge compute through the infrastructure providers it deploys on (no amounts stated).

Availability: unknown
Delivery: unknown
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=First-party pricing sources read 2026-11-10 Source-depth answers added 2026-11-16 from release main@531aedf (531aedffb672); older citations identify their own revisions.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- First-party pricing pages, terms and licences read 2026-11-10; no purchase, billing or authenticated account test.

- currency — Currency [unsupported]: No price exists. Sources: background-agents:v531-lic, background-agents:v531-readme.
- interval — Interval [unsupported]: No billing interval exists. Sources: background-agents:v531-lic, background-agents:v531-readme.
- seats — Seats [unsupported]: No seat charge exists. Sources: background-agents:v531-lic, background-agents:v531-readme.
- seat_minimum — Seat minimum [unsupported]: No seat minimum exists. Sources: background-agents:v531-lic, background-agents:v531-readme.
- workspace_fees — Workspace fees [unsupported]: No workspace fee exists. Sources: background-agents:v531-lic, background-agents:v531-readme.
- included_usage — Included usage [unsupported]: No usage is included; the operator's provider accounts bill it. Sources: background-agents:v531-lic, background-agents:v531-readme.
- credits — Credits [unsupported]: No credits are offered. Sources: background-agents:v531-lic, background-agents:v531-readme.
- trial — Trial [unsupported]: No trial exists. Sources: background-agents:v531-lic, background-agents:v531-readme.
- overages — Overages [unsupported]: No overages exist. Sources: background-agents:v531-lic, background-agents:v531-readme.
- taxes — Taxes [unsupported]: No tax treatment applies to free software. Sources: background-agents:v531-lic, background-agents:v531-readme.
- model_cost — Model cost [supported]: Models are used via the user's own API key or existing Claude/ChatGPT/SuperGrok subscriptions, billed by those providers. Sources: background-agents:readme-nov10.
- compute_cost — Compute cost [partial]: Sandboxes and control plane run on third-party platforms (Modal, Daytona, Vercel Sandbox, E2B, Boat, Cloudflare); their costs are not stated. Sources: background-agents:readme-nov10.

Checked 2026-11-10 (12 days old on 2026-11-22); agent cross-checked 2026-11-17.

Sources: background-agents:homepage-nov10, background-agents:license-nov10, background-agents:readme-nov10, background-agents:v531-lic, background-agents:v531-readme

### OWN-05: Scenario cost and controls

Verification status: 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

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Public repository, changelog and setup guides provide support context without a response commitment.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane, GitHub, Slack, Linear, API and MCP integrations; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.

- release_stage — Release stage [unknown]: Availability and support, release_stage: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- access_restrictions — Access restrictions [unknown]: Availability and support, access_restrictions: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- support — Support [partial]: Public repository, changelog and setup guides provide support context without a response commitment. Sources: background-agents:readme.
- response_commitments — Response commitments [unknown]: Availability and support, response_commitments: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- release_activity — Release activity [unknown]: Availability and support, release_activity: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- status_history — Status history [unknown]: Availability and support, status_history: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- deprecation — Deprecation [unknown]: Availability and support, deprecation: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- export_options — Export options [unknown]: Availability and support, export_options: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-08 (14 days old on 2026-11-22); agent cross-checked 2026-11-08.

Sources: background-agents:readme

### OWN-07: Enterprise assurance

Verification status: 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

Verification status: Agent cross-checked (agent-cross-checked)
Support: partial

Open-Inspect is free MIT software that operators deploy themselves; the site shows no pricing. Sandbox, hosting and model costs are paid to those providers.

Availability: generally-available
Delivery: native
Scope: surfaces=Open-Inspect Next.js web workspace and deployed control plane; version=Repository 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3; first-party source pass 2026-11-08. Pinned single-tenant source deployment. Team visibility defaults to shadow, while private/session-action checks have separate enforcement.; plan=not specified; os=not specified; execution_modes=not specified; region=not specified; harness=not specified.

Limitations:
- Self-hosted only; infrastructure and model provider costs were not checked. No installed-product test was run.

- ongoing_free — Ongoing free [supported]: The software is MIT-licensed and free to deploy; the website shows no price. Sources: background-agents:src1-license-pinned, background-agents:homepage-nov10, background-agents:setup.
- trial — Trial [unknown]: Useful free tier and evaluation access, trial: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- credits_trial — Credits trial [unknown]: Useful free tier and evaluation access, credits_trial: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- self_host_cost — Self host cost [partial]: Operators supply the control plane host, sandbox provider and model credentials; those costs are separate and not stated. Sources: background-agents:setup, background-agents:src1-readme-pinned.
- paid_only — Paid only [unsupported]: Not paid-only: the MIT source is deployed free of charge. Sources: background-agents:src1-license-pinned, background-agents:homepage-nov10.
- early_access — Early access [unknown]: Useful free tier and evaluation access, early_access: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- free_scenarios — Free scenarios [unknown]: Useful free tier and evaluation access, free_scenarios: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- card_required — Card required [unknown]: Useful free tier and evaluation access, card_required: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- usage_caps — Usage caps [unknown]: Useful free tier and evaluation access, usage_caps: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.
- paid_evaluation — Paid evaluation [unknown]: Useful free tier and evaluation access, paid_evaluation: Open-Inspect's checked sources did not establish this requirement or its rubric scenario. No installed-product fixture was run.

Checked 2026-11-10 (12 days old on 2026-11-22); agent cross-checked 2026-11-10.

Sources: background-agents:src1-license-pinned, background-agents:homepage-nov10, background-agents:setup, background-agents:src1-readme-pinned

### OWN-09: Pricing transparency

Verification status: 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.

## Sources

### background-agents:readme

Open-Inspect README.md. Self-deployed background coding system has web/chat/Git integrations, multiplayer prompting, per-session sandboxes and subtask spawning. Designed for one trusted organization, not tenant isolation.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/README.md
- Locator: Overview; Security; Architecture
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-08

### background-agents:collaboration

Open-Inspect packages/docs/content/docs/sessions/multiplayer-and-sharing.mdx. Authorized different people can prompt/stop the same session, share its timeline/sandbox and queue follow-ups. Commit attribution follows prompt authors. Team read isolation requires enforcement on; default is shadow.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/packages/docs/content/docs/sessions/multiplayer-and-sharing.mdx
- Locator: Visibility; Presence; Attribution; Patterns
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-08

### background-agents:review

Open-Inspect packages/docs/content/docs/sessions/reviewing-changes.mdx. Session Changes view displays committed/uncommitted diffs, file filters and exact comparison context. Timeline retains screenshots, recordings and command output; PR review is a separate GitHub action.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/packages/docs/content/docs/sessions/reviewing-changes.mdx
- Locator: Changes; Artifacts; PR review
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-08

### background-agents:gap35-clients

Open-Inspect package.json. The scoped user client is the web workspace. Complete package/release targets contain no native desktop or iOS/Android client. Repository command scripts provision/administer deployment, images and users; there is no product-owned coding-session CLI client in the shipped command set.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/package.json
- Locator: Workspace package list and distribution targets; packages/web; package scripts; README setup and integrations
- Evidence kind: source-inspected
- Source date: Not stated
- Retrieved: 2026-11-13
- Commit: 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3
- Path: package.json

### background-agents:harness

Open-Inspect packages/docs/content/docs/models/agent-harnesses.mdx. Sessions choose OpenCode, Claude Agent SDK or Codex for their lifetime. Children inherit harness. Claude subscriptions require Claude Agent; model choice is not native Grok CLI support.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/packages/docs/content/docs/models/agent-harnesses.mdx
- Locator: Comparison; One harness per session
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-08

### background-agents:accounts

Open-Inspect packages/docs/content/docs/models/provider-accounts.mdx. Connected Claude, ChatGPT/Codex and SuperGrok accounts or API credentials have provider defaults, identity/status and unattended policy. Subscription routes have harness-specific restrictions.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/packages/docs/content/docs/models/provider-accounts.mdx
- Locator: Connect accounts; Automated authentication
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-08

### background-agents:src3-claude-binary

Open-Inspect Claude Agent harness, credential flow. The Claude Agent harness launches the claude binary inside the sandbox through a clean-credential wrapper that passes a single Anthropic credential, either an API key or a connected Claude account token.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/docs/CLAUDE_AGENT.md#L188-L196
- Locator: docs/CLAUDE_AGENT.md lines 188-196 (runtime credential steps), commit 8749aa0
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-10

### background-agents:src3-claude-bundled

Open-Inspect Claude Agent harness behaviour. The sandbox image pins claude-agent-sdk, whose wheel bundles the claude binary; sessions run on that bundled Claude Code with its skills, Agent sub-agents and settings, plus Open-Inspect tools served over MCP.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/docs/CLAUDE_AGENT.md#L284-L321
- Locator: docs/CLAUDE_AGENT.md lines 284-321 (behaviour notes: skills, privacy, tools, sub-agents, image), commit 8749aa0
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-10

### background-agents:setup

Open-Inspect docs/SETUP_GUIDE.md. The setup guide offers running the web app against an existing backend, local contribution, or a full self-hosted deployment, and links a guide for running the control plane in a container without Cloudflare. GitHub App repository credentials are required, and GitHub/Google sign-in is distinct from them. Deployment is single-tenant.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/docs/SETUP_GUIDE.md
- Locator: Deployment paths; Path C; Important context; related guides
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-08

### background-agents:budget

Open-Inspect packages/docs/content/docs/sessions/spend-limits.mdx. Reported model spend has workspace/repo/environment/session limits. Exhaustion stops work and blocks prompts until owner raises/removes the limit. Subscription costs are informational estimates; unreported usage cannot be bounded.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/packages/docs/content/docs/sessions/spend-limits.mdx
- Locator: Measurement; Budget exhaustion
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-08

### background-agents:mcp

Open-Inspect packages/docs/content/docs/integrations/mcp-server.mdx. Locally built MCP server reads sessions/diffs/automation history and optionally creates/triggers automation/imports skills through scoped PATs, bounded by the users role.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/packages/docs/content/docs/integrations/mcp-server.mdx
- Locator: Tools; Token permissions
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-08

### background-agents:data

Open-Inspect packages/docs/content/docs/administration/data-controls.mdx. Web supports archiving/restoration, not deletion controls. Operator/service NDJSON export includes optional traces/events/usage but is best-effort, not a consistent backup.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/packages/docs/content/docs/administration/data-controls.mdx
- Locator: Archive; Delete; Session export
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-08

### background-agents:auth

Open-Inspect docs/AUTH.md. Workspace role checks, private collaborator grants and current team membership govern operations. GitHub credentials are App/session-scoped; snapshot-cached credentials and issued-token revocation limits remain documented.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/docs/AUTH.md
- Locator: Session access; Teams; Credentials; Suspension
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-08

### background-agents:tools

Open-Inspect packages/docs/content/docs/configure/mcp-servers.mdx. Managed MCP connections have shared/personal scopes and authenticated control-plane proxying. Harness-specific loading restrictions remain.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/packages/docs/content/docs/configure/mcp-servers.mdx
- Locator: Shared and personal connections
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-08

### background-agents:children

Open-Inspect packages/docs/content/docs/sessions/child-sessions.mdx. Product tools spawn independent child sandboxes/branches, send follow-ups, inspect responses/events and cancel descendants. Default limits are five concurrent and fifteen total children per parent; depth is bounded.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/packages/docs/content/docs/sessions/child-sessions.mdx
- Locator: Tools; Inheritance; Limits; UI
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-08

### background-agents:memory

Open-Inspect packages/docs/content/docs/configure/memories.mdx. Repository/session memory stores provide scoped durable records and explicit mutation/search controls; memory is separate from task-spec approval.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/packages/docs/content/docs/configure/memories.mdx
- Locator: Memory scope; Storage; Agent operations
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-08

### background-agents:gap35-native-issues

Open-Inspect terraform/d1/migrations/0002_create_session_index.sql. The scoped canonical schemas and control routes persist coding sessions, automations, memories, collaborators and external tracker references. Native inbox/search/child-task entries are session records; no standalone issue/task CRUD entity exists. Linear/GitHub remain external trackers.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/terraform/d1/migrations/0002_create_session_index.sql
- Locator: Session index and later migrations; session CRUD routes; sessions/inbox-and-search.mdx; searched independent issue/task IDs, tables and APIs across control-plane/shared code
- Evidence kind: source-inspected
- Source date: Not stated
- Retrieved: 2026-11-13
- Commit: 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3
- Path: terraform/d1/migrations/0002_create_session_index.sql

### background-agents:automation

Open-Inspect packages/docs/content/docs/automations/index.mdx. Server automation starts sessions from cron, inbound webhook, Sentry, Slack or deployed GitHub bot events. Three consecutive failures pause it; Linear automation events are planned, not current.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/packages/docs/content/docs/automations/index.mdx
- Locator: Triggers; Authorization; Failure handling
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-08

### background-agents:sandbox

Open-Inspect packages/docs/content/docs/configure/sandbox-providers.mdx. Operator selects Modal, modal-vm, Daytona, Vercel, E2B or Boat. Provider-specific snapshots/resume, CPU/RAM and timeout limits differ. Boat is a persistent full-VM sandbox; default Modal is not Docker-capable.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/packages/docs/content/docs/configure/sandbox-providers.mdx
- Locator: Comparison; Recovery models
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-08

### background-agents:gap35-execution

Open-Inspect packages/docs/content/docs/configure/sandbox-providers.mdx. The complete supported agent execution backend list is Modal, modal-vm, Daytona, Vercel, E2B and Boat. Operators configure/pay their own provider accounts; the control plane can be self-hosted locally but agent sessions execute in those remote sandboxes, not on the browser laptop. No Open-Inspect-vendor-operated compute service is offered.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/packages/docs/content/docs/configure/sandbox-providers.mdx
- Locator: Provider selection/matrix; setup provider credentials; packages/control-plane sandbox provider factory and shared provider IDs
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-13

### background-agents:runtime

Open-Inspect runtime architecture. The default development environment starts as a container; modal-vm and Boat are separately documented VM-backed providers. Runtime snapshots and host reconnection depend on provider.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/docs/HOW_IT_WORKS.md
- Locator: Sandbox boot and provider execution
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-08

### background-agents:gap35-checkouts

Open-Inspect packages/sandbox-runtime/src/sandbox_runtime/repository_boot.py. Each sandbox boots/clones its own repository checkout and branch. Child sessions receive a separate sandbox/branch. The scoped repository boot and synchronization paths use full checkouts rather than creating Git worktrees with git worktree add.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/packages/sandbox-runtime/src/sandbox_runtime/repository_boot.py
- Locator: RepositoryBoot and repository_sync.py; checkout/clone/reset operations; child session branch configuration; bounded worktree search
- Evidence kind: source-inspected
- Source date: Not stated
- Retrieved: 2026-11-13
- Commit: 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3
- Path: packages/sandbox-runtime/src/sandbox_runtime/repository_boot.py

### background-agents:limits

Open-Inspect packages/docs/content/docs/reference/limits.mdx. Documented session/child/sandbox and automation ceilings vary by configuration/provider; numeric limits are not a tested fleet capacity claim.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/packages/docs/content/docs/reference/limits.mdx
- Locator: Sessions; Boot and sandbox; Automations; MCP
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-08

### background-agents:roles

Open-Inspect packages/docs/content/docs/administration/workspace-access.mdx. Owner, Administrator, Member and Viewer have named permissions; private/team session action checks add resource-specific access boundaries. Viewer lacks prompting/sandbox access.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/packages/docs/content/docs/administration/workspace-access.mdx
- Locator: Admission; Built-in roles; Resource checks
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-08

### background-agents:gap35-auth

Open-Inspect packages/control-plane/src/auth/user/better-auth.ts. UserAuthConfig accepts GitHub and Google social OAuth providers. The complete createUserAuth provider/plugin configuration does not include enterprise SAML/OIDC SSO; workspace roles apply after sign-in. Better Auth own telemetry is explicitly disabled.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/packages/control-plane/src/auth/user/better-auth.ts
- Locator: UserAuthConfig; createUserAuth plugins/socialProviders; auth/user sign-in provider enums and workspace configuration inspected
- Evidence kind: source-inspected
- Source date: Not stated
- Retrieved: 2026-11-13
- Commit: 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3
- Path: packages/control-plane/src/auth/user/better-auth.ts

### background-agents:audit

Open-Inspect packages/docs/content/docs/administration/audit-log.mdx. Owner/admin read-only event UI distinguishes applied/no-change/denied/rejected operation outcomes from allowed authorization decisions, with actor/resource snapshots and structured details.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/packages/docs/content/docs/administration/audit-log.mdx
- Locator: Event fields; Recorded operations; Limits
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-08

### background-agents:license

Open-Inspect LICENSE. Open-Inspect repository code is MIT licensed.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/LICENSE
- Locator: MIT grant
- Evidence kind: source-inspected
- Source date: Not stated
- Retrieved: 2026-11-08
- Commit: 8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3
- Path: LICENSE

### background-agents:v531-n-telemetry

background-agents packages/ at main@531aedf. Searched telemetry, posthog, sentry, analytics, opt-out and crash report across packages/, terraform/, scripts/ and docs: matches are the in-product workspace analytics dashboard (analytics.read permission), a Sentry alert automation trigger, and a Cloudflare observability log query script; no code sends usage data to the project maintainers and no telemetry opt-out exists.

- URL: https://github.com/open-inspect/background-agents/tree/531aedffb672a9aebdc9cace7ceb310c6437adb8/packages
- Locator: bounded scan for vendor telemetry (main@531aedf, 531aedffb672)
- Evidence kind: source-inspected
- Source date: 2026-11-13
- Retrieved: 2026-11-16
- Commit: 531aedffb672a9aebdc9cace7ceb310c6437adb8
- Path: packages

### background-agents:homepage-nov10

Background Agents homepage. The site presents Open-Inspect as open source and self-hostable, with no pricing or paid plan shown.

- URL: https://backgroundagents.dev/
- Locator: Closing CTA: 'Open source. Self-hostable. Ready for your whole team.'
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-10

### background-agents:readme-nov10

Open-Inspect README. Users pick a model via their own Anthropic API key or Claude/ChatGPT/SuperGrok subscriptions, and sandboxes run on third-party infrastructure (Modal, Daytona, Vercel Sandbox, E2B, Boat, Cloudflare); no price for Open-Inspect itself is given.

- URL: https://github.com/open-inspect/background-agents/blob/main/README.md
- Locator: Overview; Multi-Provider Model Support; Credits
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-10

### background-agents:license-nov10

background-agents LICENSE. The MIT license grants permission to use, copy, modify and distribute the software free of charge.

- URL: https://github.com/open-inspect/background-agents/blob/main/LICENSE
- Locator: MIT License, permission grant
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-10

### background-agents:v531-lic

background-agents LICENSE at main@531aedf. MIT license for the repository.

- URL: https://github.com/open-inspect/background-agents/blob/531aedffb672a9aebdc9cace7ceb310c6437adb8/LICENSE#L1-L3
- Locator: MIT License header (main@531aedf, 531aedffb672)
- Evidence kind: source-inspected
- Source date: 2026-11-13
- Retrieved: 2026-11-16
- Commit: 531aedffb672a9aebdc9cace7ceb310c6437adb8
- Path: LICENSE

### background-agents:v531-readme

background-agents README.md at main@531aedf. Open-Inspect is deployed by the operator into their own Cloudflare and sandbox provider accounts; no hosted plan or price is offered.

- URL: https://github.com/open-inspect/background-agents/blob/531aedffb672a9aebdc9cace7ceb310c6437adb8/README.md#L1-L60
- Locator: overview and deployment (main@531aedf, 531aedffb672)
- Evidence kind: documented
- Source date: 2026-11-13
- Retrieved: 2026-11-16

### background-agents:src1-license-pinned

background-agents LICENSE. The MIT license grants permission to use, copy, modify and distribute the software free of charge.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/LICENSE
- Locator: LICENSE at commit 8749aa0; MIT License, permission grant
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-10

### background-agents:src1-readme-pinned

Open-Inspect README. Users pick a model via their own Anthropic API key or Claude/ChatGPT/SuperGrok subscriptions, and sandboxes run on third-party infrastructure (Modal, Daytona, Vercel Sandbox, E2B, Boat, Cloudflare); no price for Open-Inspect itself is given.

- URL: https://github.com/open-inspect/background-agents/blob/8749aa015cf92ff1caeaf64ae17ab8c7fa8938b3/README.md
- Locator: README.md at commit 8749aa0; Overview; Multi-Provider Model Support; Credits
- Evidence kind: documented
- Source date: Not stated
- Retrieved: 2026-11-10
