ci: add ClawSweeper dispatch workflow

[steipete]

Adds the standard ClawSweeper dispatch workflow so this repo's issues and PRs are reviewed by ClawSweeper (org-wide rollout, see openclaw/clawsweeper#875).

[clawsweeper]

Codex review: needs real behavior proof before merge. Reviewed July 26, 2026, 9:59 PM ET / July 27, 2026, 01:59 UTC.

ClawSweeper review

What this changes

The PR adds a 202-line GitHub Actions workflow that sends selected issue, comment, and pull-request events to ClawSweeper for automated review.

Merge readiness

⛔ Blocked until real behavior proof is added - 4 items remain

Keep open: this is a current, cleanly mergeable CI integration PR, but it adds a cross-repository dispatch and GitHub App token boundary without after-fix live-run evidence. The supplied patch follows several safe workflow practices, including a pinned token action and no checkout of untrusted PR code, so there is no discrete correctness finding; maintainers should still require a redacted proof run before merge.

Priority: P3
Reviewed head: c10a3e3954bfd5180ed3c988bade1b5c93669968

Review scores

Measure	Result	What it means
Overall readiness	🧂 unranked krab (1/6)	The implementation has useful, bounded automation scope, but missing real behavior proof prevents a merge-ready rating for a cross-repository dispatch workflow.
Proof confidence	🧂 unranked krab (1/6)	Needs real behavior proof before merge: The PR body and supplied discussion provide no after-fix live dispatch evidence; add a redacted workflow run, logs, or visible result after exercising an eligible event, and redact private endpoints, keys, IP addresses, and non-public metadata. After adding proof, update the PR body; ClawSweeper should re-review automatically. If it does not, the PR author or someone with repository write access can comment @clawsweeper re-review.
Patch quality	🦐 gold shrimp (3/6)	No actionable review findings were identified.

Verification

Check	Result	Evidence
Real behavior	Needs proof	Needs real behavior proof before merge: The PR body and supplied discussion provide no after-fix live dispatch evidence; add a redacted workflow run, logs, or visible result after exercising an eligible event, and redact private endpoints, keys, IP addresses, and non-public metadata. After adding proof, update the PR body; ClawSweeper should re-review automatically. If it does not, the PR author or someone with repository write access can comment @clawsweeper re-review.
Evidence reviewed	4 items	PR workflow surface: The supplied pull-file diff adds only .github/workflows/clawsweeper-dispatch.yml, with 202 added lines and no deletions; it introduces issue, comment, and pull_request_target event handling. Security-oriented workflow structure: The supplied diff documents that pull_request_target is used for maintainer-owned external dispatch without checkout or untrusted PR-code execution, pins actions/create-github-app-token to a full commit SHA, and gates token creation on the configured private-key secret. Current rollout status: GitHub context reports the branch is cleanly mergeable and all listed CI, lint, test, CodeQL, and secret-scanning checks completed successfully; this verifies the submitted YAML is accepted by existing repository automation but does not prove the new dispatch path ran after deployment.
Findings	None	None.
Security	None	None.

How this fits together

The repository’s GitHub event stream feeds a dispatch workflow, which filters eligible issue and pull-request activity before sending review work to the central ClawSweeper service. The resulting review automation can update the originating GitHub item while the workflow itself remains in this repository’s CI surface.

flowchart LR
  A[Issue and PR events] --> B[Dispatch workflow]
  C[Comment and bot filters] --> B
  B --> D[GitHub App token]
  D --> E[ClawSweeper service]
  E --> F[Review result on source item]

Loading

Before merge

• Add real behavior proof - Needs real behavior proof before merge: The PR body and supplied discussion provide no after-fix live dispatch evidence; add a redacted workflow run, logs, or visible result after exercising an eligible event, and redact private endpoints, keys, IP addresses, and non-public metadata. After adding proof, update the PR body; ClawSweeper should re-review automatically. If it does not, the PR author or someone with repository write access can comment @clawsweeper re-review.
• Resolve merge risk (P1) - Merging adds a pull_request_target-triggered automation boundary that can mint a GitHub App token scoped to the central ClawSweeper repository; maintainers should confirm the App installation and repository permissions remain narrowly scoped.
• Resolve merge risk (P1) - A green YAML/CI result does not show that an eligible live event reaches ClawSweeper, is filtered correctly, and cannot create dispatch loops or unwanted event volume.
• Complete next step (P2) - A maintainer should verify the privileged cross-repository dispatch boundary and require live behavior proof; this is not a mechanical code repair for ClawSweeper to perform.

Agent review details

Security

None.

Review metrics

Metric	Value	Why it matters
Workflow surface	1 workflow added; 202 lines added; 0 lines removed	The entire PR changes one privileged automation entry point, so its event and credential behavior should be reviewed as one security-sensitive unit.

Merge-risk options

Maintainer options:

1. Capture a controlled dispatch run (recommended)
   Before merging, provide redacted evidence from an eligible issue or pull-request event showing the central dispatch succeeds and a filtered event does not invoke it.
2. Accept the rollout risk explicitly
   A maintainer may merge based on the approved organization-wide workflow pattern after confirming the GitHub App installation is restricted to the intended repository and permissions.

Technical review

Best possible solution:

Land the standard dispatch workflow only after a redacted live event run demonstrates an eligible item reaches ClawSweeper and an ineligible or bot-generated event is safely ignored, while preserving the no-checkout and narrowly scoped token design.

Do we have a high-confidence way to reproduce the issue?

Not applicable as a bug report. The workflow’s intended event paths are visible in the patch, but this review has no live after-merge dispatch trace proving the new automation behavior.

Is this the best way to solve the issue?

Unclear. The workflow appears to be the standard organization-wide rollout implementation and includes sensible safeguards, but the best solution requires live evidence that the filter and central dispatch operate as intended before this privileged automation is accepted.

AGENTS.md: unclear because the file could not be read completely.

Codex review notes: model internal, reasoning high; reviewed against decf9d652ea9.

Labels

Label changes:

• add P3: This is an organization-wide CI automation rollout rather than a user-facing regression or release-blocking defect.
• add merge-risk: 🚨 security-boundary: The workflow handles untrusted GitHub event metadata and conditionally creates a GitHub App token for cross-repository dispatch.
• add merge-risk: 🚨 automation: The new workflow changes how issue, comment, and pull-request events trigger automated review work.
• add rating: 🧂 unranked krab: Overall readiness is 🧂 unranked krab; proof is 🧂 unranked krab and patch quality is 🦐 gold shrimp.
• add status: 📣 needs proof: The PR needs real behavior proof before ClawSweeper can clear the contributor ask. Needs real behavior proof before merge: The PR body and supplied discussion provide no after-fix live dispatch evidence; add a redacted workflow run, logs, or visible result after exercising an eligible event, and redact private endpoints, keys, IP addresses, and non-public metadata. After adding proof, update the PR body; ClawSweeper should re-review automatically. If it does not, the PR author or someone with repository write access can comment @clawsweeper re-review.

Label justifications:

• P3: This is an organization-wide CI automation rollout rather than a user-facing regression or release-blocking defect.
• merge-risk: 🚨 security-boundary: The workflow handles untrusted GitHub event metadata and conditionally creates a GitHub App token for cross-repository dispatch.
• merge-risk: 🚨 automation: The new workflow changes how issue, comment, and pull-request events trigger automated review work.
• rating: 🧂 unranked krab: Overall readiness is 🧂 unranked krab; proof is 🧂 unranked krab and patch quality is 🦐 gold shrimp.
• status: 📣 needs proof: The PR needs real behavior proof before ClawSweeper can clear the contributor ask. Needs real behavior proof before merge: The PR body and supplied discussion provide no after-fix live dispatch evidence; add a redacted workflow run, logs, or visible result after exercising an eligible event, and redact private endpoints, keys, IP addresses, and non-public metadata. After adding proof, update the PR body; ClawSweeper should re-review automatically. If it does not, the PR author or someone with repository write access can comment @clawsweeper re-review.

Evidence

What I checked:

• PR workflow surface: The supplied pull-file diff adds only .github/workflows/clawsweeper-dispatch.yml, with 202 added lines and no deletions; it introduces issue, comment, and pull_request_target event handling. (.github/workflows/clawsweeper-dispatch.yml:1, c10a3e3954bf)
• Security-oriented workflow structure: The supplied diff documents that pull_request_target is used for maintainer-owned external dispatch without checkout or untrusted PR-code execution, pins actions/create-github-app-token to a full commit SHA, and gates token creation on the configured private-key secret. (.github/workflows/clawsweeper-dispatch.yml:10, c10a3e3954bf)
• Current rollout status: GitHub context reports the branch is cleanly mergeable and all listed CI, lint, test, CodeQL, and secret-scanning checks completed successfully; this verifies the submitted YAML is accepted by existing repository automation but does not prove the new dispatch path ran after deployment. (.github/workflows/clawsweeper-dispatch.yml:1, c10a3e3954bf)
• Repository policy status: The local inspection sandbox could not execute read commands, so a full target-repository AGENTS.md read and local history provenance check could not be confirmed in this review.

Likely related people:

• steipete: Authored the sole workflow commit and linked the change to the organization-wide ClawSweeper rollout context; no separate current-main workflow owner could be established from the available read-only evidence. (role: automation rollout contributor; confidence: low; commits: c10a3e3954bf; files: .github/workflows/clawsweeper-dispatch.yml)

Rank-up moves

Optional improvements that raise the rating; they are not merge blockers.

• Post a redacted live workflow trace showing a qualifying event dispatches to ClawSweeper and produces the expected review result.
• Show one filtered bot or non-command event does not dispatch, without exposing credentials or private event payloads.

Rating scale

Score	Internal tier	Crab rank	Meaning
6/6	S	🦀 challenger crab	Exceptional readiness
5/6	A	🦞 diamond lobster	Very strong readiness
4/6	B	🐚 platinum hermit	Good normal PR; ordinary maintainer review
3/6	C	🦐 gold shrimp	Useful, but confidence is limited
2/6	D	🦪 silver shellfish	Proof or implementation needs work
1/6	F	🧂 unranked krab	Not merge-ready
N/A	NA	🌊 off-meta tidepool	Rating does not apply

Overall follows the weaker of proof and patch quality.
Shiny media proof means a screenshot, video, or linked artifact directly shows the changed behavior. Runtime, network, CSP, and security claims still need visible diagnostics.

Workflow

• ClawSweeper keeps one durable marker-backed review comment per issue or PR.
• Re-runs edit this comment so the latest verdict, findings, and automation markers stay together instead of adding duplicate bot comments.
• A fresh review can be triggered by eligible @clawsweeper re-review comments, exact-item GitHub events, scheduled/background review runs, or manual workflow dispatch.
• PR/issue authors and users with repository write access can comment @clawsweeper re-review or @clawsweeper re-run on an open PR or issue to request a fresh review only.
• Maintainers can also comment @clawsweeper review to request a fresh review only.
• Fresh-review commands do not start repair, autofix, rebase, CI repair, or automerge.
• Maintainer-only repair and merge flows require explicit commands such as @clawsweeper autofix, @clawsweeper automerge, @clawsweeper fix ci, or @clawsweeper address review.
• Maintainers can comment @clawsweeper explain to ask for more context, or @clawsweeper stop to stop active automation.