Auto Review Loop (MiniMax Version): 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.md for legacy projects)
• REVIEWER_MODEL = MiniMax-M2.7 — Model used via MiniMax API

API Configuration

This skill uses MiniMax API for external review. Two methods are supported:

Method 1: MCP Tool (Primary)

If mcp__minimax-chat__minimax_chat is available, use it:

mcp__minimax-chat__minimax_chat:
  message: |
    [Review prompt content]
  model: "MiniMax-M2.7"
  system: "You are a senior machine learning researcher..."

Method 2: curl (Fallback)

If MCP is not available, use curl directly:

curl -s "https://api.minimax.io/v1/chat/completions" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $MINIMAX_API_KEY" \
  -d '{
    "model": "MiniMax-M2.7",
    "messages": [
      {"role": "system", "content": "You are a senior ML researcher..."},
      {"role": "user", "content": "[Review prompt]"}
    ],
    "max_tokens": 4096
  }'

API Key: Read from ~/.codex/settings.json under env.MINIMAX_API_KEY, or from environment variable.

Why MiniMax instead of a secondary Codex agent? Codex CLI uses OpenAI's Responses API (/v1/responses) which is not supported by third-party providers. See: https://github.com/openai/codex/discussions/7782

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,
  "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

1. Check for review-stage/REVIEW_STATE.json (fall back to ./REVIEW_STATE.json if not found — legacy path):
   • If neither path exists: fresh start (normal case)
   • If it exists AND status is "completed": fresh start (previous loop finished normally)
   • If it exists AND status is "in_progress" AND timestamp is older than 24 hours: fresh start (stale state from a killed/abandoned run — delete the file and start over)
   • If it exists AND status is "in_progress" AND timestamp is within 24 hours: resume
     • Read the state file to recover round, last_score, pending_experiments
     • Read review-stage/AUTO_REVIEW.md to restore full context of prior rounds (fall back to ./AUTO_REVIEW.md)
     • If pending_experiments is 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."
2. Read project narrative documents, memory files, and any prior review documents
3. Read recent experiment results (check output directories, logs)
4. Identify current weaknesses and open TODOs from prior reviews
5. Initialize round counter = 1 (unless recovered from state file)
6. Create/update review-stage/AUTO_REVIEW.md with header and timestamp

Loop (repeat up to MAX_ROUNDS)

Phase A: Review

Send comprehensive context to the external reviewer.

Check MCP availability first, then use appropriate method:

If MCP available (Primary):

Use mcp__minimax-chat__minimax_chat tool with:
- system: "You are a senior machine learning researcher serving as a reviewer for top-tier conferences like NeurIPS, ICML, and ICLR. Provide rigorous, constructive feedback."
- prompt: [Full review prompt with context]
- model: "MiniMax-M2.7"

If MCP NOT available (Fallback):

curl -s "https://api.minimax.io/v1/chat/completions" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $MINIMAX_API_KEY" \
  -d '{
    "model": "MiniMax-M2.7",
    "messages": [
      {
        "role": "system",
        "content": "You are a senior machine learning researcher serving as a reviewer for top-tier conferences like NeurIPS, ICML, and ICLR. Provide rigorous, constructive feedback."
      },
      {
        "role": "user",
        "content": "[Round N/MAX_ROUNDS of autonomous review loop]\n\n[Full research context: claims, methods, results, known weaknesses]\n[Changes since last round, if any]\n[For round 2+: Summary of previous review feedback and what was addressed]\n\nPlease act as a senior ML reviewer (NeurIPS/ICML level).\n\n1. Score this work 1-10 for a top venue\n2. List remaining critical weaknesses (ranked by severity)\n3. For each weakness, specify the MINIMUM fix (experiment, analysis, or reframing)\n4. State clearly: is this READY for submission? Yes/No/Almost\n\nBe brutally honest. If the work is ready, say so clearly."
      }
    ],
    "max_tokens": 4096
  }'

Note: Each round is a standalone API call. For round 2+, include the summary of previous reviews and changes in the prompt itself.

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.

Phase C: Implement Fixes (if not stopping)

For each action item (highest priority first):

1. Code changes: Write/modify experiment scripts, model code, analysis scripts
2. Run experiments: Deploy to GPU server via SSH + screen/tmux
3. Analysis: Run evaluation, collect results, update figures/tables
4. 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, score, verdict, and any pending experiments.

Increment round counter → back to Phase A.

Termination

When loop ends (positive assessment or max rounds):

1. Update review-stage/REVIEW_STATE.json with "status": "completed"
2. Write final sum