Auto Review Loop LLM

Autonomous research review loop using any OpenAI-compatible LLM API. Configure via llm-chat MCP server or environment variables. Trigger with "auto review loop llm" or "llm review".

Published by @wanshuiyin·0 agent reads / 30d·0 saves·

Auto Review Loop (Generic LLM): Autonomous Research Improvement

🔒 Do not wrap this skill in /loop, /schedule, or CronCreate. Like /auto-review-loop, it already loops internally (review → fix → re-review), feeding each round's prior-round summary into the next review prompt (the backend is a stateless per-round API/MCP call, not a shared thread). An external timer re-enters from the top each tick, dropping that accumulated context and firing the verdict on wall-clock time instead of on artifact change — zero new signal, full token cost. Schedule the external wait that precedes it, not the verdict. See shared-references/external-cadence.md.

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 check below. Verdict vocabulary is {"ready", "almost", "not ready"}. (Earlier wording used or and a stale verdict set; the AND form is authoritative.)
  • REVIEW_DOC: review-stage/AUTO_REVIEW.md (cumulative log) (fall back to ./AUTO_REVIEW.md for legacy projects)

LLM Configuration

This skill uses any OpenAI-compatible API for external review via the llm-chat MCP server.

Configuration via MCP Server (Recommended)

Add to ~/.claude/settings.json:

{
  "mcpServers": {
    "llm-chat": {
      "command": "/usr/bin/python3",
      "args": ["/Users/yourname/.claude/mcp-servers/llm-chat/server.py"],
      "env": {
        "LLM_API_KEY": "your-api-key",
        "LLM_BASE_URL": "https://api.deepseek.com/v1",
        "LLM_MODEL": "deepseek-chat"
      }
    }
  }
}

Supported Providers

ProviderLLM_BASE_URLLLM_MODEL
OpenAIhttps://api.openai.com/v1gpt-4o, o3
DeepSeekhttps://api.deepseek.com/v1deepseek-chat, deepseek-reasoner
MiniMaxhttps://api.minimax.io/v1MiniMax-M3
Kimi (Moonshot)https://api.moonshot.cn/v1moonshot-v1-8k, moonshot-v1-32k
ZhiPu (GLM)https://open.bigmodel.cn/api/paas/v4glm-4, glm-4-plus
SiliconFlowhttps://api.siliconflow.cn/v1Qwen/Qwen2.5-72B-Instruct
阿里云百炼https://dashscope.aliyuncs.com/compatible-mode/v1qwen-max
零一万物https://api.lingyiwanwu.com/v1yi-large

API Call Method

Primary: MCP Tool

mcp__llm-chat__chat:
  prompt: |
    [Review prompt content]
  model: "deepseek-chat"
  system: "You are a senior ML reviewer..."

Fallback: curl

curl -s "${LLM_BASE_URL}/chat/completions" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer ${LLM_API_KEY}" \
  -d '{
    "model": "${LLM_MODEL}",
    "messages": [
      {"role": "system", "content": "You are a senior ML reviewer..."},
      {"role": "user", "content": "[review prompt]"}
    ],
    "max_tokens": 4096
  }'

State Persistence (Compact Recovery)

Persist state to review-stage/REVIEW_STATE.json after each round:

{
  "round": 2,
  "status": "in_progress",
  "last_score": 5.0,
  "last_verdict": "not ready",
  "pending_experiments": [],
  "timestamp": "2026-03-15T10:00:00"
}

Write this file at the end of every Phase E (after documenting the round).

On completion, set "status": "completed".

Workflow

Initialization

  1. Check review-stage/REVIEW_STATE.json for recovery (fall back to ./REVIEW_STATE.json if not found — legacy path)
  2. Read project context and prior reviews
  3. Initialize round counter

Loop (up to MAX_ROUNDS)

Phase A: Review

If MCP available:

mcp__llm-chat__chat:
  system: "You are a senior ML reviewer (NeurIPS/ICML level)."
  prompt: |
    [Round N/MAX_ROUNDS of autonomous review loop]

    [Full research context: claims, methods, results, known weaknesses]
    [Changes since last round, if any]

    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
    4. State clearly: is this READY for submission? Yes/No/Almost

    Be brutally honest. If the work is ready, say so clearly.

If MCP NOT available:

curl -s "${LLM_BASE_URL}/chat/completions" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer ${LLM_API_KEY}" \
  -d '{
    "model": "${LLM_MODEL}",
    "messages": [
      {"role": "system", "content": "You are a senior ML reviewer (NeurIPS/ICML level)."},
      {"role": "user", "content": "[Full review prompt]"}
    ],
    "max_tokens": 4096
  }'
Phase B: Parse Assessment

CRITICAL: Save the FULL raw response verbatim. Then extract:

  • Score (numeric 1-10)
  • Verdict ("ready" / "almost" / "not ready")
  • Action items (ranked list of fixes)

STOP: If score >= 6 AND verdict ∈ {"ready", "almost"} (exact — "not ready" does NOT qualify)

Phase C: Implement Fixes

Priority: metric additions > reframing > new experiments

Phase D: Wait for Results

Monitor remote experiments

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 here — verbatim, unedited.]

</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 state.

Termination

  1. Set review-stage/REVIEW_STATE.json status to "completed"
  2. Write final summary

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.

  • Anti-hallucination citations: When adding references, NEVER fabricate BibTeX. Use DBLP → CrossRef → [VERIFY] chain. Do NOT generate BibTeX from memory.

  • Be honest about weaknesses

  • Implement fixes BEFORE re-reviewing

  • Document everything

  • Include previous context in round 2+ prompts

  • Prefer MCP tool over curl when available

Prompt Template for Round 2+

mcp__llm-chat__chat:
  system: "You are a senior ML reviewer (NeurIPS/ICML level)."
  prompt: |
    [Round N/MAX_ROUNDS of autonomous review loop]

    ## Previous Review Summary (Round N-1)
    - Previous Score: X/10
    - Previous Verdict: [ready/almost/not ready]
    - Previous Key Weaknesses: [list]

    ## Changes Since Last Review
    1. [Action 1]: [result]
    2. [Action 2]: [result]

    ## Updated Results
    [paste updated metrics/tables]

    Please re-score and re-assess:
    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
    4. State clearly: is this READY for submission? Yes/No/Almost

    Be brutally honest. If the work is ready, say so clearly.

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 MANIFEST.md
  • Output Language Protocol — respect the project's language setting

Bundled with this artifact

4 files

Reference files that ship alongside this artifact. Agents pull these in only when the task needs them.

More on the bench

SKILL0

Writing Systems Papers

Paragraph-level structural blueprint for 10-12 page systems papers targeting OSDI, SOSP, ASPLOS, NSDI, and EuroSys. Provides page allocation, paragraph templates, and writing patterns. Use when user says "写系统论文", "systems paper structure", "OSDI paper", "SOSP paper", or wants fine-grained structural guidance for a systems conference submission.

ai-prompt-engineering+1
0
SKILL0

Wiki Enrich

Fill in the per-paper TODO sections of research-wiki/papers/<slug>.md pages that literature-ingest skills leave as bare scaffolds. Use when user says 'enrich wiki', 'fill paper TODOs', 'wiki body 補完', '把 paper 摘要寫進 wiki', 'research-wiki 自動填', or after a batch ingest that left papers/ as TODO scaffolds.

ai-prompt-engineering+1
0
SKILL0

Vast Gpu

Rent, manage, and destroy GPU instances on vast.ai. Use when user says "rent gpu", "vast.ai", "rent a server", "cloud gpu", or needs on-demand GPU without owning hardware.

ai-prompt-engineering+1
0