ClawSweeper documentation
- Status: active documentation index
- Owner: ClawSweeper maintainers
- Source of truth: the linked repository files and their owning code,
- Last verified:
openclaw/clawsweeper@647503ec44b8e777dd172adf974a945367da0d19 - Update when: a document is added, retired, moved, changes lifecycle, or gains a
configuration, workflows, and tests
new canonical owner
ClawSweeper reviews GitHub issues and pull requests, publishes durable results, and applies or repairs only through guarded maintainer-controlled paths. Start with the task or audience below; this page is a map, not another operator manual.
#Start here
| Goal | First page | Next step |
|---|---|---|
| Understand the product boundary | Project vision | Orchestration |
| Set up a development checkout | Contributing | Local run |
| Understand review and scheduling | Scheduler | Automation limits |
| Add or operate a target repository | Target repositories | Target dispatcher |
| Inspect production without mutation | Live dashboard | OpenClaw Bay |
| Inspect observer routes and configuration ownership | Public observer API | Operator configuration |
| Operate repair or automerge | Repair entry point | Repair operations |
| Prepare proof for a change | Contributing | Agent instructions |
| Review a local committed range | Local branch review | pnpm local-review -- --base origin/main |
Production mutation commands are intentionally not repeated here. Follow the linked runbook, confirm its status and source of truth, and keep all execution, fix, merge, automerge, deployment, and queue gates closed unless the current operator has explicit authority to open them.
#Document lifecycle
Every evergreen page belongs to one of these states:
- active: current guidance for a supported surface
- proposed: a future design or runbook that grants no execution approval
- compatibility-only: retained for still-readable state or a supported
- historical: completed proof or decision history, not current procedure
migration boundary, but not for creating new work
Active runbooks and volatile configuration references should begin with status, role owner, owning source, last-verified revision, and update triggers. “Owner” is a responsibility role until maintainers approve a stable GitHub CODEOWNERS individual or team. Generated or checked values should name their checker; everything else requires human comparison with current main.
#Concepts and architecture
- Project vision — active; product goals and boundaries
- Orchestration — active; planner, review, publish, apply,
- Scheduler — active and volatile; scheduled and exact-event
- State storage — active; Worker, R2, and Git-backed state
- Action ledger — active; mutation receipts and replay
- Work lane — active; bounded workflow admission
- Review cache — active; reusable review-state contract
- GitHub webhook read model — active; signed
and repair shape
execution behavior
ownership
boundaries
ingress, App subscriptions, snapshot freshness, repair, and safety boundary
#Configuration and repository onboarding
- Automation limits — active; validated by
- Target repositories — active; profiles, generic
- Target dispatcher — active; exact-event integration
- OpenClaw event hooks — active; event routes
- Local branch review — active local/GitHub-isolated
pnpm run check:limits
fallbacks, inventory, and apply membership
replacement for hosted commit review
#Operations and runbooks
- Live dashboard — active; public observer surface and
- Queue-service split runbook — proposed;
- Repair operations — active; canonical repair operator
- Local ClawSweeper skill — active; read-only
operator diagnostics
unapproved production migration plan
procedures, trust checks, gates, and token boundaries
local exact-item or committed-range review
#Dashboards and observability
- Live dashboard — active; Worker status and operational
- GitHub publication egress telemetry — active;
- OpenClaw Bay — active; public six-lane visualization
- Triage dashboard — active; cached issue triage
- PR proof triage — active; maintainer proof
telemetry
wire denominator, credential attribution, completeness, and retention
inspection
#Contributor, review, and proof guidance
- Contributing — active; setup, scope, PR evidence, and
- Agent instructions — active and binding for coding agents
- PR review comments — active; review-thread handling
- Related issue discovery — active; duplicate and
- Live proof — active; secretless browser/terminal execution,
author-owned review loop
adjacent-report context
trusted media attachment, and local simulation
docs/proof/** contains inspectable artifacts for specific changes. Those files support their recorded claim but do not override current runbooks.
#Proof media retention
Commit lightweight, diffable proof records: the claim and limits, exact source or head SHA, reproduction command, provider/image/lease provenance, fixture hashes, observable assertions, and compact normalized JSON. Full-resolution screenshots, videos, and traces may be committed while a pull request is under review, but prune them from main after merge once the durable review and proof record exist; git history retains the review-time bytes.
Keep a heavy binary only when an active canonical document embeds or links it, or when runtime or test tooling consumes it. Prefer one current compressed poster or WebP over accumulated before/after sets.
#Repair and automerge
- Repair entry point — active; concepts, task map, and local
- Repair operations — active; canonical operator
- Internal feature map — active implementation
- Steerable repair architecture — active;
- Trusted PR repair — active; autofix/automerge
- Automatic issue PRs — active; guarded issue
- Automerge flow — active; state machine
- Automerge proof — active; controlled proof harness
- Containment validation history —
command reference
procedures
reference, not an operator runbook
end-to-end protocol and control boundaries
contract
implementation intake
historical; completed July 2026 work
#Close policy
- Implemented-on-main paired close policy
- Obsolescence policies — active policy map
- Unsponsored feature policy — active
- Product direction policy — active
- Author PR budget policy — active
- Stalled PR policies — active
— active policy and formal GitHub-link requirement
#Security and special-purpose surfaces
- Spam scanner — active; trust and classification boundary
- Security-sensitive work must use maintainer security handling rather than a
normal public PR or public documentation page.
#Update triggers
Review the relevant pages in the same change when any of these surfaces move:
config/automation-limits.jsonor workflow limit literals: automation limitsconfig/target-repositories.json, repository profiles, target inventory, or- exact-review queue schema, retry classification, publication capacity,
- dashboard route, projection, or public-field changes: dashboard and Bay docs
- publication GitHub request paths, credential selection, or egress fields:
- mutation, close, proof, repair, or merge policy changes: policy pages,
- workflow retirement or replacement: every command example and compatibility
and scheduler
apply membership: target-repository and dispatcher docs
batching, direct publication, or state writing: scheduler, limits, dashboard, Bay, and queue runbooks
GitHub egress telemetry and public API docs
CONTRIBUTING, AGENTS, and repair operations
page that names the workflow
#Automated drift checks
pnpm run check:docs validates exact-case relative links and Markdown anchors, documented package scripts, workflow files, repository source paths, and the selected configuration-derived claims in config/documentation-sync.json. It runs through the existing static check and CI path. Add a manifest claim when volatile production configuration is quoted as a concrete value in prose. Stage newly added files before running the check so its repository inventory matches the proposed commit rather than unrelated untracked workspace contents.
The check deliberately does not crawl external URLs, infer policy, assign ownership, or treat historical proof commands and paths as current contracts. The index, lifecycle labels, and whether a configuration claim should become policy remain human-reviewed.
The public documentation build reads config/documentation-site.json as its exhaustive lifecycle manifest. Active pages appear in navigation, search, the sitemap, and llms.txt. Proposed, historical, and proof pages remain available at stable generated paths for evidence links, but carry a visible lifecycle banner and noindex metadata and are excluded from canonical discovery. The documentation check fails when a page is unclassified or classified twice.