ClawHub PR Maintainer
Use this skill for maintainer-facing ClawHub GitHub workflow, not for ordinary
implementation work.
Start With Live GitHub State
- Use
gh pr view or gh issue view against openclaw/clawhub; verify live
state before commenting, labeling, closing, or recommending merge.
- For PRs, read title, body, author, labels, comments, files, commits, status
checks, review state, and linked issues.
- Surface author identity briefly: GitHub name/login and account age when
useful. Treat identity as triage signal, never as proof by itself.
Common read-only commands:
gh pr view <number> --repo openclaw/clawhub --json title,body,author,labels,comments,files,commits,statusCheckRollup,reviewDecision,url,additions,deletions,changedFiles
gh issue view <number> --repo openclaw/clawhub --json title,body,author,labels,comments,state,url
gh api users/<login> --jq '{login,name,created_at,type}'
Review Evidence Bar
- For bug fixes, require symptom evidence, a plausible root cause in the touched
code path, and either a regression test or focused manual proof.
- For UI changes, require screenshots or video when the behavior is meaningfully
visual. Use tests as supplemental evidence, not a substitute for visible proof.
- Do not merge or recommend merge based only on PR prose, AI rationale, or green
CI when the changed behavior has not been exercised.
- For contributor-provided screenshots/videos/logs, inspect the artifact
directly and state what it proves. Do not rerun
proof:ui just to inspect
existing evidence.
Structure PR Review Output
- Start every PR review with 1-3 plain sentences explaining what the change does
and why it matters.
- Show size near the top as
LOC: +x/-y (N files), using live PR stats or
local diff stats.
- Then list findings first. If none, say
No blocking findings or
No findings.
- Always answer: affected ClawHub surface, bug or behavior being changed,
evidence checked, and best-fix verdict.
- For bug/regression fixes, include a compact
Provenance: line when a bounded
history pass identifies it. Separate code author, PR author,
merger/committer, current PR author, PR number, and date when those differ.
If the blamed PR was merged by automation, identify the human trigger when
practical; otherwise say trigger unknown.
Read Beyond The Diff
- For code-path bug, regression, or behavior changes, review the surrounding
path, not just changed lines. Open the runtime entry point, owner module, one
caller, one callee, adjacent tests, and sibling surfaces that should share the
invariant.
- For docs/config/process-only changes, read the changed file, its linked or
adjacent source of truth, and any route/workflow/template the change claims to
affect. Do not require runtime caller/callee evidence when no runtime path
exists.
- Compare against current
origin/main behavior or current published docs when
regression, compatibility, or user-visible docs accuracy matters.
- For dependency-backed behavior, read the upstream docs/source/types before
judging API use, defaults, output shapes, errors, timeouts, memory behavior, or
compatibility.
- Mention the main files or contracts read when the verdict depends on
code-path, docs, config, or workflow evidence.
- If a required path is uninspected, keep reading or mark
Remaining uncertainty; do not call the PR best, proof-sufficient, or
merge-ready.
Best-Fix Review Loop
Every PR review must explicitly answer: "Is this the best fix, or only a
plausible fix?"
Before verdict:
- Reconstruct the bug, feature need, or behavior claim from the issue, PR, and
proof.
- For code-path changes, trace current behavior from entry point to failure or
decision point.
- For docs/config/process-only changes, trace the reader/operator workflow or
automation path the change is meant to clarify.
- Read touched files, relevant callers/callees for code changes, adjacent docs
or tests, owner modules, and relevant source-of-truth docs.
- Read sibling surfaces that should share the invariant or could be broken by a
one-sided fix.
- Compare against current
origin/main and shipped behavior when relevant.
- Identify at least one alternative fix location or shape, then reject it with
evidence.
Review output must include:
Best-fix verdict: best / acceptable mitigation / wrong layer / too narrow /
too broad.
Alternatives considered: 1-3 concrete alternatives and why rejected.
Code read: compact list of main files/contracts checked.
Remaining uncertainty: what was not proven.
Enforce Bug-Fix Evidence
- Never merge a bug-fix PR based only on issue text, PR text, or AI rationale.
- Before recommending merge for a bug fix, require:
- symptom evidence such as a repro, logs, failing test, or focused manual
proof
- a verified root cause in code with file/line
- blame-backed provenance for regressions when traceable, or commit SHA/date
when no PR is traceable
- a fix that touches the implicated code path
- a regression test when feasible, or explicit manual verification plus a
reason no test was added
- If the claim is unsubstantiated or likely wrong, request evidence or changes
instead of recommending merge.
Decide UI Proof Mode
Use the clawhub-ui-proof skill when the maintainer/agent should generate new
visual evidence.
before-after: bug fixes, regressions, changed copy, changed layout, or any
PR where main-vs-candidate comparison clarifies the change.
feature: new page, new flow, new UI state, or behavior that cannot exist on
origin/main.
- No generated proof: docs-only, backend-only, tests-only, metadata-only, or
already-sufficient contributor evidence.
Write a temporary Playwright scenario under .artifacts/proof-scenarios/; do
not infer manual clicks. Keep screenshots and videos in .artifacts/ until
publishing. Never commit proof artifacts.
Final Review Comment With Proof
If this review generated proof:ui artifacts, publish them before the final PR
review comment. Do not leave only local .artifacts/... paths in a PR comment;
they are useful to the maintainer locally but invisible to GitHub readers.
Use:
bun run proof:publish -- --proof-dir .artifacts/clawhub-ui-proof/<timestamp> --target-pr <number>
proof:publish copies the selected files to the qa-artifacts branch and
upserts a marker-backed PR comment with a ClawHub UI Proof section.
That comment includes:
- the proof mode (
before-after or feature)
- the
report.md result summary
- the most relevant per-step screenshots
- inline video previews when GIF previews are present
- links to full-run MP4s
- links to raw proof files on the artifact branch
Use --dry-run before publishing if you need to inspect the generated comment.
If publishing fails because credentials are missing, report the local proof
directory and the failed command instead of posting a comment that claims
evidence is attached.
ClawSweeper
ClawSweeper is the bot control plane for automated PR/issue review once ClawHub
dispatch is configured. Until then, use this skill for manual maintainer review.
If ClawSweeper has posted a review, read it as evidence but verify live PR state
before acting.
Commenting And Labels
- Use literal multiline comment bodies or
--body-file; never pass escaped
\n strings.
- For issue comments and PR comments containing backticks or shell characters,
prefer a single-quoted heredoc or
--body-file over inline -b bodies.
- Do not wrap issue or PR refs like
#123 in backticks when you want GitHub to
auto-link them.
- Keep maintainer comments short: finding, evidence, requested action, and
verification path.
- When no proof artifacts were generated,
gh pr comment --body-file is fine.
When proof artifacts were generated, use proof:publish so screenshots/videos
are published before posting.
- Do not close more than five issues/PRs in one action without explicit
confirmation and the exact target list.
1---2name: clawhub-pr-maintainer3description: Use when reviewing, triaging, validating, or discussing ClawHub GitHub issues or pull requests, including author context, CI, UI proof, evidence, labels, close decisions, and maintainer handoff.4metadata:5 source: https://skillmd.com/skills/openclaw/clawhub-pr-maintainer6---78# ClawHub PR Maintainer910Use this skill for maintainer-facing ClawHub GitHub workflow, not for ordinary11implementation work.1213## Start With Live GitHub State1415- Use `gh pr view` or `gh issue view` against `openclaw/clawhub`; verify live16 state before commenting, labeling, closing, or recommending merge.17- For PRs, read title, body, author, labels, comments, files, commits, status18 checks, review state, and linked issues.19- Surface author identity briefly: GitHub name/login and account age when20 useful. Treat identity as triage signal, never as proof by itself.2122Common read-only commands:2324```sh25gh pr view <number> --repo openclaw/clawhub --json title,body,author,labels,comments,files,commits,statusCheckRollup,reviewDecision,url,additions,deletions,changedFiles26gh issue view <number> --repo openclaw/clawhub --json title,body,author,labels,comments,state,url27gh api users/<login> --jq '{login,name,created_at,type}'28```2930## Review Evidence Bar3132- For bug fixes, require symptom evidence, a plausible root cause in the touched33 code path, and either a regression test or focused manual proof.34- For UI changes, require screenshots or video when the behavior is meaningfully35 visual. Use tests as supplemental evidence, not a substitute for visible proof.36- Do not merge or recommend merge based only on PR prose, AI rationale, or green37 CI when the changed behavior has not been exercised.38- For contributor-provided screenshots/videos/logs, inspect the artifact39 directly and state what it proves. Do not rerun `proof:ui` just to inspect40 existing evidence.4142## Structure PR Review Output4344- Start every PR review with 1-3 plain sentences explaining what the change does45 and why it matters.46- Show size near the top as `LOC: +x/-y (N files)`, using live PR stats or47 local diff stats.48- Then list findings first. If none, say `No blocking findings` or49 `No findings`.50- Always answer: affected ClawHub surface, bug or behavior being changed,51 evidence checked, and best-fix verdict.52- For bug/regression fixes, include a compact `Provenance:` line when a bounded53 history pass identifies it. Separate code author, PR author,54 merger/committer, current PR author, PR number, and date when those differ.55 If the blamed PR was merged by automation, identify the human trigger when56 practical; otherwise say trigger unknown.5758## Read Beyond The Diff5960- For code-path bug, regression, or behavior changes, review the surrounding61 path, not just changed lines. Open the runtime entry point, owner module, one62 caller, one callee, adjacent tests, and sibling surfaces that should share the63 invariant.64- For docs/config/process-only changes, read the changed file, its linked or65 adjacent source of truth, and any route/workflow/template the change claims to66 affect. Do not require runtime caller/callee evidence when no runtime path67 exists.68- Compare against current `origin/main` behavior or current published docs when69 regression, compatibility, or user-visible docs accuracy matters.70- For dependency-backed behavior, read the upstream docs/source/types before71 judging API use, defaults, output shapes, errors, timeouts, memory behavior, or72 compatibility.73- Mention the main files or contracts read when the verdict depends on74 code-path, docs, config, or workflow evidence.75- If a required path is uninspected, keep reading or mark76 `Remaining uncertainty`; do not call the PR best, proof-sufficient, or77 merge-ready.7879## Best-Fix Review Loop8081Every PR review must explicitly answer: "Is this the best fix, or only a82plausible fix?"8384Before verdict:85861. Reconstruct the bug, feature need, or behavior claim from the issue, PR, and87 proof.882. For code-path changes, trace current behavior from entry point to failure or89 decision point.903. For docs/config/process-only changes, trace the reader/operator workflow or91 automation path the change is meant to clarify.924. Read touched files, relevant callers/callees for code changes, adjacent docs93 or tests, owner modules, and relevant source-of-truth docs.945. Read sibling surfaces that should share the invariant or could be broken by a95 one-sided fix.966. Compare against current `origin/main` and shipped behavior when relevant.977. Identify at least one alternative fix location or shape, then reject it with98 evidence.99100Review output must include:101102- `Best-fix verdict:` best / acceptable mitigation / wrong layer / too narrow /103 too broad.104- `Alternatives considered:` 1-3 concrete alternatives and why rejected.105- `Code read:` compact list of main files/contracts checked.106- `Remaining uncertainty:` what was not proven.107108## Enforce Bug-Fix Evidence109110- Never merge a bug-fix PR based only on issue text, PR text, or AI rationale.111- Before recommending merge for a bug fix, require:112 1. symptom evidence such as a repro, logs, failing test, or focused manual113 proof114 2. a verified root cause in code with file/line115 3. blame-backed provenance for regressions when traceable, or commit SHA/date116 when no PR is traceable117 4. a fix that touches the implicated code path118 5. a regression test when feasible, or explicit manual verification plus a119 reason no test was added120- If the claim is unsubstantiated or likely wrong, request evidence or changes121 instead of recommending merge.122123## Decide UI Proof Mode124125Use the `clawhub-ui-proof` skill when the maintainer/agent should generate new126visual evidence.127128- `before-after`: bug fixes, regressions, changed copy, changed layout, or any129 PR where main-vs-candidate comparison clarifies the change.130- `feature`: new page, new flow, new UI state, or behavior that cannot exist on131 `origin/main`.132- No generated proof: docs-only, backend-only, tests-only, metadata-only, or133 already-sufficient contributor evidence.134135Write a temporary Playwright scenario under `.artifacts/proof-scenarios/`; do136not infer manual clicks. Keep screenshots and videos in `.artifacts/` until137publishing. Never commit proof artifacts.138139## Final Review Comment With Proof140141If this review generated `proof:ui` artifacts, publish them before the final PR142review comment. Do not leave only local `.artifacts/...` paths in a PR comment;143they are useful to the maintainer locally but invisible to GitHub readers.144145Use:146147```sh148bun run proof:publish -- --proof-dir .artifacts/clawhub-ui-proof/<timestamp> --target-pr <number>149```150151`proof:publish` copies the selected files to the `qa-artifacts` branch and152upserts a marker-backed PR comment with a **ClawHub UI Proof** section.153154That comment includes:155156- the proof mode (`before-after` or `feature`)157- the `report.md` result summary158- the most relevant per-step screenshots159- inline video previews when GIF previews are present160- links to full-run MP4s161- links to raw proof files on the artifact branch162163Use `--dry-run` before publishing if you need to inspect the generated comment.164If publishing fails because credentials are missing, report the local proof165directory and the failed command instead of posting a comment that claims166evidence is attached.167168## ClawSweeper169170ClawSweeper is the bot control plane for automated PR/issue review once ClawHub171dispatch is configured. Until then, use this skill for manual maintainer review.172If ClawSweeper has posted a review, read it as evidence but verify live PR state173before acting.174175## Commenting And Labels176177- Use literal multiline comment bodies or `--body-file`; never pass escaped178 `\n` strings.179- For issue comments and PR comments containing backticks or shell characters,180 prefer a single-quoted heredoc or `--body-file` over inline `-b` bodies.181- Do not wrap issue or PR refs like `#123` in backticks when you want GitHub to182 auto-link them.183- Keep maintainer comments short: finding, evidence, requested action, and184 verification path.185- When no proof artifacts were generated, `gh pr comment --body-file` is fine.186 When proof artifacts were generated, use `proof:publish` so screenshots/videos187 are published before posting.188- Do not close more than five issues/PRs in one action without explicit189 confirmation and the exact target list.