mohammadhprp/teammate/review-work
Claude Code and agent skill by mohammadhprp/teammate
- community
Checking that the SKILL.md file is still in the repository.
Coordinate review of delegated work: decide who reviews a returned worker's result, obtain an independent review or apply the shared review-change method yourself, record the verdict, and route a fail to run-rework. Use when a worker returns and its result needs checking, or when a rework round came back for another pass.
Install
npx vskill@latest install mohammadhprp/teammate/review-workSKILL.md
The same SKILL.md was published earlier as mohammadhprp/teammate/review-work.
Review work
Completion is not correctness, and you own the work you delegated. This skill coordinates review — who reviews, what verdict is recorded, and what happens when the work fails. The review method itself (evidence, findings, severity, verdict rules) is shared with every other reviewer, so apply review-change instead of re-deriving it here.
When to use
- A worker has returned and its result must be checked against the acceptance
criteria.
review_policyinteam-mate.tomlrequires review, or the change is
important or risky enough that the implementer should not judge it alone.
- Rework came back and needs another pass.
When not to use
- The task is still running: see
monitor-agents. - You only need the review method (findings, severity, verdict): that is
review-change.
Inputs
- The returned task from the ledger (
python3 scripts/tm.py task show <id>): its
goal, acceptance criteria, and iteration.
- The worker's fresh output —
report,diff, and the project's own check
commands (from load-project-context).
review_policyfromteam-mate.toml, and thereview-changemethod the
reviewer applies.
Procedure
- Assemble the review inputs. Read the persisted task with
python3 scripts/tm.py task show <id> for the goal, criteria, and iteration, then collect the worker's current output so the reviewer has fresh material:
``bash python3 scripts/tm.py report "<name>" python3 scripts/tm.py diff --cwd "<root>" --stat ``
- Choose the reviewer. Read
review_policy: underalways, review every
returned task; under on-risk, review important or risky changes; under never, review only what the developer asks. Apply review-change yourself, and follow independent-review to add a separate reviewer worker that applies review-task when it is warranted. The reviewer gets the objective, acceptance criteria, and evidence — not the implementer's conclusion — and returns findings only.
- Record the findings, then the verdict. The findings schema, categories,
severity levels, and pass/fail rules live in review-change's references/findings.md. Record the findings as data so rework reads the ledger rather than prose, and record only the verdict status; the iteration increment belongs to run-rework:
``bash python3 scripts/tm.py task findings <id> --file <findings.json> # an object or an array python3 scripts/tm.py task update <id> --status ready_for_approval --report-file <f> # pass python3 scripts/tm.py task update <id> --status rework --report-file <f> # fail python3 scripts/tm.py task update <id> --verdict inconclusive --report-file <f> # cannot determine ``
task show renders the verdict and the open findings, so a resumed primary reads the same state the reviewer recorded. On an inconclusive verdict, record it and escalate to the developer (step 5); do not set ready_for_approval — the approval gate refuses a task whose verdict is not pass. On a pass, first close the findings that now hold — task update and task decide refuse a pass while an open blocker/major remains:
``bash python3 scripts/tm.py task resolve <id> --all # findings fixed or accepted ``
<id> is the build task: the findings and verdict belong to it, and a pass moves only that task to ready_for_approval. If the review worker is itself ledgered, its own task is --kind review and never enters the approval queue (task-ledger); a review is evidence for the build decision, not something the developer approves.
- Drive rework. On
fail, followrun-reworkto send the open findings
back and repeat monitor → review until it converges or max_iterations is reached.
- Escalate. On the iteration limit, an
inconclusiveverdict, a worker
that returned a question or failed, or a disputed finding that blocks progress, follow escalate-decision: include the full report and one clear question — the developer decides; you recommend.
Output
A recorded verdict and either a converged pass or a bounded escalation.
Failure
An implementer cannot pass its own work, and a pass with zero checks run is a process failure. When evidence is missing, the verdict is inconclusive, not pass.
Security
- Security scanPassed, 90 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingsrc/skills/review-work/SKILL.md
Popularity
- 7-day trend score
- 0
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/mohammadhprp/teammate/review-work)