auto-review-loop
Skill9 starsChanged 3 months agoMITView on GitHub
Prompt-only by default in the lightweight pack; Claude/Gemini reviewer transport is explicit opt-in. Autonomous multi-round research review loop. Repeatedly reviews using Claude Code via claude-review MCP, implements fixes, and re-reviews until positive assessment or max rounds reached. Use when user says \"auto review loop\", \"review until it passes\", or wants autonomous iterative improvement.
7 other projects use a version of this, 6 of them edited.
Adopted by
Most recently taken up by
- yvesory/academic-research-suiteedited, first seen Oct 9, 2026
- talhaMah56/claude_skillstheir own version, first seen Sep 12, 2026
- Shri-Phnx/claude-skillsedited, first seen Sep 3, 2026
- EarthOnlinePlayer5732/my-skillsedited, first seen Aug 11, 2026
- wanshuiyin/Auto-claude-code-research-in-sleepedited, first seen Jul 15, 2026
- raja21068/AutoResearchedited, first seen May 5, 2026
Other versions7 projects use a version of this.
Word for word · 0
None found.
Edited · 6
- wanshuiyin/Auto-claude-code-research-in-sleep★ 19k11% same
- EarthOnlinePlayer5732/my-skills★ 1735% same
- yvesory/academic-research-suite★ 935% same
- Yusong-Enceladus/claude-skills★ 874% same
- raja21068/AutoResearch★ 243% same
- Shri-Phnx/claude-skills★ 226% same
Wrote their own version · 1
- talhaMah56/claude_skills★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository47 other files it gives its agents.
- autodl-hpcSkill
- experiment-bridgeSkill
- experiment-planSkill
- human-toneSkill
- idea-discoverySkill
- paper-visualizationSkill
- paper-writingSkill
- rebuttalSkill
- research-blueprintSkill
- research-pipelineSkill
- research-repo-architectSkill
- research-reviewSkill
- resubmit-pipelineSkill
- autodl-hpcSkill
- auto-paper-improvement-loopSkill
- novelty-checkSkill
- paper-figureSkill
- paper-planSkill
- paper-writeSkill
- research-refineSkill
- research-reviewSkill
- experiment-bridgeSkill
- experiment-planSkill
- auto-paper-improvement-loopSkill
- auto-review-loopSkill
- grant-proposalSkill
- idea-creatorSkill
- idea-discoverySkill
- novelty-checkSkill
- paper-figureSkill
- paper-planSkill
- paper-posterSkill
- paper-slidesSkill
- paper-writeSkill
- paper-writingSkill
- research-refineSkill
- research-reviewSkill
- human-toneSkill
- idea-discoverySkill
- paper-visualizationSkill
- paper-writingSkill
- rebuttalSkill
- research-blueprintSkill
- research-pipelineSkill
- research-repo-architectSkill
- research-reviewSkill
- resubmit-pipelineSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
39 rules. 1 is common, 22 are distinctive, and 16 are very specific to this project.
1
Common
Most of the 7 edited versions we read keep these.
22
Distinctive
This author's own choices. Worth a look.
16
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 14
Write the review prompt under review-prompts/ and ask the user to run it in a separate conversation with the relevant skills enabled.
- Distinctive: line 16
Do not call Claude/Gemini reviewer MCP bridges, APIs, or follow-up status tools by default.
- Very specific: line 18
Use pasted review text as the external verdict, then write compact summaries such as REVIEW_SUMMARY.md, NEXT_ACTIONS.md, or the …
- Distinctive: line 20
Keep heavy experiments AutoDL/HPC-gated and user-approved.
- Common: line 65
If neither path exists: fresh start (normal case, identical to behavior before this feature existed)
- Very specific: line 66
If it exists AND status is "completed": fresh start (previous loop finished normally)
- Very specific: line 67
If it exists AND status is "in_progress" AND timestamp is older than 24 hours: fresh start (stale state from a killed/abandoned run — …
- Very specific: line 68
If it exists AND status is "in_progress" AND timestamp is within 24 hours: resume
- Very specific: line 69
Read the state file to recover round, thread_id, last_score, pending_experiments
- Very specific: line 70
Read review-stage/AUTO_REVIEW.md to restore full context of prior rounds *(fall back to ./AUTO_REVIEW.md)*
- Very specific: line 71
If pending_experiments is non-empty, check if they have completed (e.g., check screen sessions)
- Distinctive: line 74
Read project narrative documents, memory files, and any prior review documents
- Distinctive: line 75
Read recent experiment results (check output directories, logs)
- Very specific: line 78
Create/update review-stage/AUTO_REVIEW.md with header and timestamp
- Very specific: line 104
After this start call, immediately save the returned jobId and poll mcp__claude-review__review_status with a bounded waitSeconds until …
- Very specific: line 106
If this is round 2+, use mcp__claude-review__review_reply_start with the saved completed threadId, then poll …
- Distinctive: line 117
STOP CONDITION: If score >= 6 AND verdict ∈ {"ready", "almost"} (exact match — "not ready" does NOT qualify) → stop loop, document final …
- Very specific: line 123
When HUMAN_CHECKPOINT = true, present the review results and wait for user input:
- Distinctive: line 146
Wait for the user's response.
- Very specific: line 154
After parsing the score, check if ~/.codex/feishu.json exists and mode is not "off":
- Distinctive: line 156
If interactive mode and verdict is "almost": send as checkpoint, wait for user reply on whether to continue or stop
- Distinctive: line 157
If config absent or mode off: skip entirely (no-op)
- Distinctive: line 163
Code changes: Write/modify experiment scripts, model code, analysis scripts
- Distinctive: line 167
Analysis: Run evaluation, collect results, update figures/tables
- Distinctive: line 168
Documentation: Update project notes and review document
- Distinctive: line 173
Prefer reframing/analysis over new experiments when both address the concern
- Distinctive: line 174
Always implement metric additions (cheap, high impact)
- Distinctive: line 178
If experiments were launched:
- Distinctive: line 220
When loop ends (positive assessment or max rounds):
- Very specific: line 222
Update review-stage/REVIEW_STATE.json with "status": "completed"
- Very specific: line 223
Write final summary to review-stage/AUTO_REVIEW.md
- Distinctive: line 224
Update project notes with conclusions
- Distinctive: line 225
If stopped at max rounds without positive assessment:
- Very specific: line 240
Large file handling: If the Write tool fails due to file size, immediately retry using Bash (cat << 'EOF' > file) to write in chunks.
- Distinctive: line 242
Always ask the Claude reviewer for strict, high-rigor feedback.
- Distinctive: line 245
Do NOT hide weaknesses to game a positive score
- Distinctive: line 247
If an experiment takes > 30 minutes, launch it and continue with other fixes while waiting
- Distinctive: line 248
Document EVERYTHING — the review log should be self-contained
- Distinctive: line 249
Update project notes after each round, not just at the end
“Common” is measured: kept by at least 6 in 10 of the 7 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Override for Codex users who want Claude Code, not a second Codex agent, to act as the reviewer. Install this package after
skills/skills-codex/*.
Customized Pack Defaults
This reviewer overlay is opt-in transport only. In the customized lightweight pack, default to prompt-only review unless the user explicitly asks to use this overlay reviewer backend:
- Write the review prompt under
review-prompts/and ask the user to run it in a separate conversation with the relevant skills enabled. - Do not call Claude/Gemini reviewer MCP bridges, APIs, or follow-up status tools by default.
- Use pasted review text as the external verdict, then write compact summaries
such as
REVIEW_SUMMARY.md,NEXT_ACTIONS.md, or the skill-specific log. - Keep heavy experiments AutoDL/HPC-gated and user-approved; local work stays at edits, tests, lint, dry-runs, and tiny smoke checks.
Auto Review Loop: Autonomous Research Improvement
Autonomously iterate: review → implement fixes → re-review, until the external reviewer gives a positive assessment or MAX_ROUNDS is reached.
Context: $ARGUMENTS
Constants
- MAX_ROUNDS = 4
- POSITIVE_THRESHOLD: score >= 6/10 AND verdict ∈ {"ready", "almost"} — both must hold, matching the operative STOP CONDITION below. Verdict vocabulary is {"ready", "almost", "not ready"}. (Earlier wording used "or" + a stale verdict set; the AND form is authoritative.)
- REVIEW_DOC:
review-stage/AUTO_REVIEW.md(cumulative log) (fall back to./AUTO_REVIEW.mdfor legacy projects) - OUTPUT_DIR =
review-stage/— Directory for review output files. - REVIEWER_MODEL =
claude-review— Claude reviewer invoked through the localclaude-reviewMCP bridge. SetCLAUDE_REVIEW_MODELif you need a specific Claude model override. - HUMAN_CHECKPOINT = false — When
true, pause after each round's review (Phase B) and present the score + weaknesses to the user. Wait for user input before proceeding to Phase C. The user can: approve the suggested fixes, provide custom modification instructions, skip specific fixes, or stop the loop early. Whenfalse(default), the loop runs fully autonomously.
💡 Override:
/auto-review-loop "topic" — human checkpoint: true
State Persistence (Compact Recovery)
Long-running loops may hit the context window limit, triggering automatic compaction. To survive this, persist state to review-stage/REVIEW_STATE.json after each round:
{
"round": 2,
"thread_id": "019cd392-...",
"status": "in_progress",
"last_score": 5.0,
"last_verdict": "not ready",
"pending_experiments": ["screen_name_1"],
"timestamp": "2026-03-13T21:00:00"
}Write this file at the end of every Phase E (after documenting the round). Overwrite each time — only the latest state matters.
On completion (positive assessment or max rounds), set "status": "completed" so future invocations don't accidentally resume a finished loop.
Workflow
Initialization
- Check for
review-stage/REVIEW_STATE.json(fall back to./REVIEW_STATE.jsonif not found — legacy path):- If neither path exists: fresh start (normal case, identical to behavior before this feature existed)
- If it exists AND
statusis"completed": fresh start (previous loop finished normally) - If it exists AND
statusis"in_progress"ANDtimestampis older than 24 hours: fresh start (stale state from a killed/abandoned run — delete the file and start over) - If it exists AND
statusis"in_progress"ANDtimestampis within 24 hours: resume- Read the state file to recover
round,thread_id,last_score,pending_experiments - Read
review-stage/AUTO_REVIEW.mdto restore full context of prior rounds (fall back to./AUTO_REVIEW.md) - If
pending_experimentsis non-empty, check if they have completed (e.g., check screen sessions) - Resume from the next round (round = saved round + 1)
- Log: "Recovered from context compaction. Resuming at Round N."
- Read the state file to recover
- Read project narrative documents, memory files, and any prior review documents
- Read recent experiment results (check output directories, logs)
- Identify current weaknesses and open TODOs from prior reviews
- Initialize round counter = 1 (unless recovered from state file)
- Create/update
review-stage/AUTO_REVIEW.mdwith header and timestamp
Loop (repeat up to MAX_ROUNDS)
Phase A: Review
Send comprehensive context to the external reviewer:
mcp__claude-review__review_start:
prompt: |
[Round N/MAX_ROUNDS of autonomous review loop]
[Full research context: claims, methods, results, known weaknesses]
[Changes since last round, if any]
Please act as a senior ML reviewer (NeurIPS/ICML level).
1. Score this work 1-10 for a top venue
2. List remaining critical weaknesses (ranked by severity)
3. For each weakness, specify the MINIMUM fix (experiment, analysis, or reframing)
4. State clearly: is this READY for submission? Yes/No/Almost
Be brutally honest. If the work is ready, say so clearly.After this start call, immediately save the returned jobId and poll mcp__claude-review__review_status with a bounded waitSeconds until done=true. Treat the completed status payload's response as the reviewer output, and save the completed threadId for any follow-up round.
If this is round 2+, use mcp__claude-review__review_reply_start with the saved completed threadId, then poll mcp__claude-review__review_status with the returned jobId until done=true to maintain continuity.
Phase B: Parse Assessment
CRITICAL: Save the FULL raw response from the external reviewer verbatim (store in a variable for Phase E). Do NOT discard or summarize — the raw text is the primary record.
Then extract structured fields:
- Score (numeric 1-10)
- Verdict ("ready" / "almost" / "not ready")
- Action items (ranked list of fixes)
STOP CONDITION: If score >= 6 AND verdict ∈ {"ready", "almost"} (exact match — "not ready" does NOT qualify) → stop loop, document final state.
Human Checkpoint (if enabled)
Skip this step entirely if HUMAN_CHECKPOINT = false.
When HUMAN_CHECKPOINT = true, present the review results and wait for user input:
📋 Round N/MAX_ROUNDS review complete.
Score: X/10 — [verdict]
Top weaknesses:
1. [weakness 1]
2. [weakness 2]
3. [weakness 3]
Suggested fixes:
1. [fix 1]
2. [fix 2]
3. [fix 3]
Options:
- Reply "go" or "continue" → implement all suggested fixes
- Reply with custom instructions → implement your modifications instead
- Reply "skip 2" → skip fix #2, implement the rest
- Reply "stop" → end the loop, document current stateWait for the user's response. Parse their input:
- Approval ("go", "continue", "ok", "proceed"): proceed to Phase C with all suggested fixes
- Custom instructions (any other text): treat as additional/replacement guidance for Phase C. Merge with reviewer suggestions where appropriate
- Skip specific fixes ("skip 1,3"): remove those fixes from the action list
- Stop ("stop", "enough", "done"): terminate the loop, jump to Termination
Feishu Notification (if configured)
After parsing the score, check if ~/.codex/feishu.json exists and mode is not "off":
- Send a
review_scorednotification: "Round N: X/10 — [verdict]" with top 3 weaknesses - If interactive mode and verdict is "almost": send as checkpoint, wait for user reply on whether to continue or stop
- If config absent or mode off: skip entirely (no-op)
Phase C: Implement Fixes (if not stopping)
For each action item (highest priority first):
- Code changes: Write/modify experiment scripts, model code, analysis scripts
- Run experiments: for heavy compute, prepare an AutoDL/HPC handoff or
invoke
/experiment-bridge; do not deploy via SSH/screen/tmux unless the user explicitly approves that legacy path. - Analysis: Run evaluation, collect results, update figures/tables
- Documentation: Update project notes and review document
Prioritization rules:
- Skip fixes requiring excessive compute (flag for manual follow-up)
- Skip fixes requiring external data/models not available
- Prefer reframing/analysis over new experiments when both address the concern
- Always implement metric additions (cheap, high impact)
Phase D: Wait for Results
If experiments were launched:
- Monitor remote sessions for completion
- Collect results from output files and logs
Phase E: Document Round
Append to review-stage/AUTO_REVIEW.md:
## Round N (timestamp)
### Assessment (Summary)
- Score: X/10
- Verdict: [ready/almost/not ready]
- Key criticisms: [bullet list]
### Reviewer Raw Response
<details>
<summary>Click to expand full reviewer response</summary>
[Paste the COMPLETE raw response from the external reviewer here — verbatim, unedited.
This is the authoritative record. Do NOT truncate or paraphrase.]
</details>
### Actions Taken
- [what was implemented/changed]
### Results
- [experiment outcomes, if any]
### Status
- [continuing to round N+1 / stopping]Write review-stage/REVIEW_STATE.json with current round, agent id, score, verdict, and any pending experiments.
Increment round counter → back to Phase A.
Termination
When loop ends (positive assessment or max rounds):
- Update
review-stage/REVIEW_STATE.jsonwith"status": "completed" - Write final summary to
review-stage/AUTO_REVIEW.md - Update project notes with conclusions
- If stopped at max rounds without positive assessment:
- List remaining blockers
- Estimate effort needed for each
- Suggest whether to continue manually or pivot
- Feishu notification (if configured): Send
pipeline_donewith final score progression table
Output Protocols
Follow these shared protocols for all output files:
- Output Versioning Protocol — write timestamped file first, then copy to fixed name
- Output Manifest Protocol — log every output to docs/project/OUTPUT_MANIFEST.md
- Output Language Protocol — respect the project's language setting
Key Rules
- Large file handling: If the Write tool fails due to file size, immediately retry using Bash (
cat << 'EOF' > file) to write in chunks. Do NOT ask the user for permission — just do it silently. - Always ask the Claude reviewer for strict, high-rigor feedback.
- Save the completed
threadIdfrom the firstmcp__claude-review__review_statusresult, then usemcp__claude-review__review_reply_startplusmcp__claude-review__review_statusfor subsequent rounds - Be honest — include negative results and failed experiments
- Do NOT hide weaknesses to game a positive score
- Implement fixes BEFORE re-reviewing (don't just promise to fix)
- If an experiment takes > 30 minutes, launch it and continue with other fixes while waiting
- Document EVERYTHING — the review log should be self-contained
- Update project notes after each round, not just at the end
Prompt Template for Round 2+
mcp__claude-review__review_reply_start:
threadId: [saved from round 1]
prompt: |
[Round N update]
Since your last review, we have:
1. [Action 1]: [result]
2. [Action 2]: [result]
3. [Action 3]: [result]
Updated results table:
[paste metrics]
Please re-score and re-assess. Are the remaining concerns addressed?
Same format: Score, Verdict, Remaining Weaknesses, Minimum Fixes.After this start call, immediately save the returned jobId and poll mcp__claude-review__review_status with a bounded waitSeconds until done=true. Treat the completed status payload's response as the reviewer output, and save the completed threadId for any follow-up round.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

