agentleFS
YulivuYulivu/debuffer-skills/skills/research-review/SKILL.md

research-review

Skill9 starsChanged 3 months agoMITView on GitHub

Prepare a prompt-only deep critical review package for research ideas, papers, or results. Generates concise reviewer prompts for a separate conversation, supports venue profiles such as ICLR, AAAI, JMLR, and TPAMI, and consumes pasted feedback; direct reviewer backends are opt-in. Use when user says "review my research", "help me review", or "get external review".

18 other projects use a version of this, 11 of them edited.

Adopted by

Most recently taken up by

Used with

Projects with it also use formula-derivation, paper-illustration and 5 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • raja21068Skill

    formula-derivation

    Structures and derives research formulas when the user wants to 推导公式, build a theory line, organize assumptions, turn scattered equations into a coherent derivation, or rewrite theory notes into a paper-ready formula document. Use when the derivation target is not yet fully fixed, the main object still needs to be chosen, or the user needs a coherent derivation package rather than a finished theorem proof.

    Also in 6 projects that use this

  • raja21068Skill

    paper-illustration

    Generate publication-quality AI illustrations for academic papers using Gemini image generation. Creates architecture diagrams, method illustrations with Claude-supervised iterative refinement loop. Use when user says \"生成图表\", \"画架构图\", \"AI绘图\", \"paper illustration\", \"generate diagram\", or needs visual figures for papers.

    Also in 5 projects that use this

  • PostHogSkill

    experiment-audit

    Audit a PostHog A/B experiment for a customer — verify config, exposure, attribution, and metrics. Trigger phrases include \"audit [customer]'s experiment\", \"audit the [name] experiment\", \"check experiment setup for [customer]\", \"validate this A/B test\", or any request to review whether an experiment is correctly wired up. Assumes you already have MCP access to the customer's project (typically via the impersonation flow set up by the `impersonate-audit` wrapper that ships with this plugin).

    Also in 4 projects that use this

  • rongaredeSkill

    semantic-scholar

    使用 Semantic Scholar API 检索和验证学术论文。支持并发多关键词搜索、批量 ID/DOI 查询、引用分析、批量补全摘要、统一 Markdown 导出。覆盖 2.14 亿+ 学术论文,无需 API Key 即可使用。触发词:论文检索、论文验证、Semantic Scholar、S2 搜索、查论文、补全摘要、导出MD

    Also in 6 projects that use this

  • +3 more from raja21068

The file

Show its rules and text

We split this skill into the rules it gives an agent

17 rules. 3 are common, 3 are distinctive, and 11 are very specific to this project.

  • 3

    Common

    Most of the 16 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 11

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    REVIEWER_BACKEND = prompt-only: Do not call Codex MCP, Manual Review MCP, Oracle, or other reviewer APIs unless the user explicitly …

    line 41
  2. Very specific:

    Write review-prompts/research_review_prompt.md with concise context, target venue profile, artifact paths, claims, evidence, known …

    line 44
  3. Distinctive:

    Ask the user to open a separate conversation with this skill package enabled and paste the review back.

    line 47
  4. Very specific:

    Write compact review-stage/REVIEW_SUMMARY.md and NEXT_ACTIONS.md.

    line 52
  5. Very specific:

    When calling the reviewer, branch on REVIEWER_BACKEND:

    line 64
  6. Very specific:

    Use mcp__codex__codex for new review threads.

    line 67
  7. Very specific:

    Use mcp__codex__codex-reply for follow-up rounds (reuse threadId).

    line 68
  8. Very specific:

    Use mcp__manual_review__review for new review threads with: prompt: [exact same prompt that would go to Codex] config: …

    line 71
  9. Very specific:

    Use mcp__manual_review__review_reply for follow-up rounds with: threadId: [saved manual-review threadId] prompt: [follow-up prompt] config: …

    line 75
  10. Common:

    Before calling the external reviewer, compile a comprehensive briefing:

    line 96
  11. Common:

    Read project narrative documents (e.g., STORY.md, README.md, paper drafts)

    line 97
  12. Common:

    Read any memory/notes files for key findings and experiment history

    line 98
  13. Distinctive:

    Use the appropriate tool to continue the conversation:

    line 125
  14. Very specific:

    Update project memory/notes with key review conclusions, using docs/project/NEXT_ACTIONS.md for follow-up work when needed.

    line 154
  15. Very specific:

    ALWAYS use config: {"model_reasoning_effort": "xhigh"} for reviews

    line 159
  16. Distinctive:

    Document the threadId for potential future resumption

    line 164
  17. Very specific:

    After each reviewer call (mcp__codex__codex, mcp__codex__codex-reply, mcp__manual_review__review, or mcp__manual_review__review_reply), …

    line 186

“Common” is measured: kept by at least 6 in 10 of the 16 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Research Review via Prompt-Only External Reviewer

Capability Routing

This is a first-layer entry skill. Keep it loaded as the user-facing route; when a request needs a specialized capability below, resolve the debuffer repo root from .debuffer_skills/installed-skills-codex.txt (repo_root) when available, read the referenced library SKILL.md, then follow that skill. Do not copy the whole library skill into this file.

  • /experiment-audit: read ../library/review/experiment-audit/SKILL.md.
  • /experiment-writeup-audit: read ../library/review/experiment-writeup-audit/SKILL.md.
  • /paper-claim-audit: read ../library/review/paper-claim-audit/SKILL.md.
  • /citation-audit: read ../library/review/citation-audit/SKILL.md.
  • /proof-checker: read ../library/review/proof-checker/SKILL.md.
  • /auto-review-loop: read ../library/review/auto-review-loop/SKILL.md.
  • /final-experiment-curator: read ../library/review/final-experiment-curator/SKILL.md.
  • /submission-repo-cleanroom: read ../library/review/submission-repo-cleanroom/SKILL.md.

🔒 Do not wrap this skill in /loop, /schedule, or CronCreate. It is verdict-bearing — it produces a cross-model review verdict, multi-round with reviewer thread continuity. An external timer re-fires the verdict on wall-clock time and breaks the reviewer's round-to-round memory: zero new signal, full token cost. Schedule the external wait that precedes it (work ready → then review once), not the verdict. See shared-references/external-cadence.md.

Get a multi-round critical review of research work from the selected external reviewer backend with maximum reasoning depth.

Customized Pack Defaults

Read ../shared-references/lightweight-research-pack.md, ../shared-references/venue-profiles.md, and ../shared-references/model-policy.md before preparing the review. Default to prompt-only review:

  • REVIEWER_BACKEND = prompt-only. Do not call Codex MCP, Manual Review MCP, Oracle, or other reviewer APIs unless the user explicitly requests a legacy backend such as codex, manual, or oracle-pro.
  • Write review-prompts/research_review_prompt.md with concise context, target venue profile, artifact paths, claims, evidence, known weaknesses, and exact reviewer questions.
  • Ask the user to open a separate conversation with this skill package enabled and paste the review back. Treat the pasted review as the external verdict.
  • Preserve Review Tracing by saving the prompt, pasted review, and response summary under review-stage/, but avoid copying long reviewer text into main project docs.
  • Write compact review-stage/REVIEW_SUMMARY.md and NEXT_ACTIONS.md; only create long review reports when requested.

Constants

  • REVIEWER_MODEL = external conversation chosen by user
  • REVIEWER_BACKEND = prompt-only. Legacy backends (codex, manual, oracle-pro) are not part of the lightweight default and require an explicit user request plus available tools.

Reviewer Calling Convention

When calling the reviewer, branch on REVIEWER_BACKEND:

If REVIEWER_BACKEND = codex: Use mcp__codex__codex for new review threads. Use mcp__codex__codex-reply for follow-up rounds (reuse threadId).

If REVIEWER_BACKEND = manual: Use mcp__manual_review__review for new review threads with: prompt: [exact same prompt that would go to Codex] config: {"model_reasoning_effort": "xhigh"} Save the returned threadId. Use mcp__manual_review__review_reply for follow-up rounds with: threadId: [saved manual-review threadId] prompt: [follow-up prompt] config: {"model_reasoning_effort": "xhigh"}

Prompt fidelity: the manual prompt must be exactly the same text that Codex would receive. Review tracing applies equally to both backends.

Context: $ARGUMENTS

Prerequisites

Prompt-only review has no provider prerequisite. For an explicit automatic backend, verify that the requested adapter is installed and record the resolved model, effort, and thread/task handle in the review trace. If the adapter is unavailable, stop with a blocked result instead of silently switching reviewers.

Workflow

Step 1: Gather Research Context

Before calling the external reviewer, compile a comprehensive briefing:

  1. Read project narrative documents (e.g., STORY.md, README.md, paper drafts)
  2. Read any memory/notes files for key findings and experiment history
  3. Identify: core claims, methodology, key results, known weaknesses

Step 2: Initial Review (Round 1)

Send a detailed prompt using the selected backend and its supported effort setting. The prompt-only path is the default.

For codex backend:

mcp__codex__codex:
  config: {"model_reasoning_effort": "xhigh"}
  prompt: |
    [Full research context + specific questions]
    Please act as a senior ML reviewer (NeurIPS/ICML level). Identify:
    1. Logical gaps or unjustified claims
    2. Missing experiments that would strengthen the story
    3. Narrative weaknesses
    4. Whether the contribution is sufficient for a top venue
    Please be brutally honest.

For manual backend: use mcp__manual_review__review with the same prompt and config: {"model_reasoning_effort": "xhigh"}. Save the returned threadId.

Step 3: Iterative Dialogue (Rounds 2-N)

For codex backend: use mcp__codex__codex-reply with the returned threadId. For manual backend: use mcp__manual_review__review_reply with the same threadId. Use the appropriate tool to continue the conversation:

For each round:

  1. Respond to criticisms with evidence/counterarguments
  2. Ask targeted follow-ups on the most actionable points
  3. Request specific deliverables: experiment designs, paper outlines, claims matrices

Key follow-up patterns:

  • "If we reframe X as Y, does that change your assessment?"
  • "What's the minimum experiment to satisfy concern Z?"
  • "Please design the minimal additional experiment package (highest acceptance lift per GPU week)"
  • "Please write a mock NeurIPS/ICML review with scores"
  • "Give me a results-to-claims matrix for possible experimental outcomes"

Step 4: Convergence

Stop iterating when:

  • Both sides agree on the core claims and their evidence requirements
  • A concrete experiment plan is established
  • The narrative structure is settled

Step 5: Document Everything

Save the full interaction and conclusions to a review document under review-stage/:

  • Round-by-round summary of criticisms and responses
  • Final consensus on claims, narrative, and experiments
  • Claims matrix (what claims are allowed under each possible outcome)
  • Prioritized TODO list with estimated compute costs
  • Paper outline if discussed

Update project memory/notes with key review conclusions, using docs/project/NEXT_ACTIONS.md for follow-up work when needed.

Key Rules

  • ALWAYS use config: {"model_reasoning_effort": "xhigh"} for reviews
  • Send comprehensive context in Round 1 — the external model cannot read your files
  • Be honest about weaknesses — hiding them leads to worse feedback
  • Push back on criticisms you disagree with, but accept valid ones
  • Focus on ACTIONABLE feedback — "what experiment would fix this?"
  • Document the threadId for potential future resumption
  • The review document should be self-contained (readable without the conversation)

Prompt Templates

For initial review:

"I'm going to present a complete ML research project for your critical review. Please act as a senior ML reviewer (NeurIPS/ICML level)..."

For experiment design:

"Please design the minimal additional experiment package that gives the highest acceptance lift per GPU week. Our compute: [describe]. Be very specific about configurations."

For paper structure:

"Please turn this into a concrete paper outline with section-by-section claims and figure plan."

For claims matrix:

"Please give me a results-to-claims matrix: what claim is allowed under each possible outcome of experiments X and Y?"

For mock review:

"Please write a mock NeurIPS review with: Summary, Strengths, Weaknesses, Questions for Authors, Score, Confidence, and What Would Move Toward Accept."

Review Tracing

After each reviewer call (mcp__codex__codex, mcp__codex__codex-reply, mcp__manual_review__review, or mcp__manual_review__review_reply), save the trace following shared-references/review-tracing.md (Policy C — forensic; never silently skip). Use save_trace.sh (resolved per the chain in shared-references/integration-contract.md §2) or write files directly to .debuffer_skills/traces/<skill>/<date>_run<NN>/. Respect the --- trace: parameter (default: full).

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.