How do I install the Co Review skill?

Run npx skillmds@latest add zy-ning/co-review in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Co Review safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Co Review?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Co Review free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Co Review?

zy-ning (@zy-ning) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Co Review by zy-ning — Agent Skill

Name: Co Review
Author: zy-ning

Co Review

Adversarial critique of a research draft or paper using an isolated critic context. The critic must have no access to the current conversation history — context isolation, not model identity, is what makes the review independent. Use when a draft is complete or partially complete, after each writing round, or when `co-write` invokes it automatically. Returns FATAL/MAJOR/MINOR issues, a structured score (0–10), and a PROCEED/REFINE/PIVOT recommendation. Caps at 4 rounds. Trigger phrases: "review paper", "critique draft", "check citations", "adversarial review".

zy-ning 391e691 2 files · 4.2 KB Updated Sep 17, 2026

File contents

Co-Review

Run an adversarial review in an isolated context. The critic must have no access to the current conversation history — this is what makes the critique independent, not which model is used.

Critic Selection

Use the first available isolated critic in priority order:
- Claude subagent (via Task tool or similar) with a clean context — preferred when no external MCP is configured, same model is fine
- codex:codex (Codex MCP) — isolated by the MCP boundary
- auto-review-loop-llm skill (any OpenAI-compatible endpoint)
- auto-review-loop-minimax skill (MiniMax)
Record which critic was used in RESEARCH.md Context alongside the score.
If no isolated context can be established, say so and stop. Do not review within the same conversation context.

Review Rubric (Fixed — Do Not Modify)

Send the work to the critic with this exact rubric:

VERIFIABLE (graded on evidence, not opinion):

Every claim maps to an experiment result in RESEARCH.md Context. Flag claims that don't.
No hallucinated citations. Check every DOI and arXiv ID. Flag any that don't resolve.
Ablations support stated conclusions. Check the numbers match.

SUBJECTIVE (graded by the critic):

Clarity of problem statement (1–5)
Novelty relative to cited work (1–5)
Writing quality (1–5)

The critic must return a structured result matching the schema in RUBRIC.md. See RUBRIC.md for field definitions, the PROCEED/REFINE/PIVOT decision rules, and the verbatim integrity instruction to include in every critic prompt.

Integrity Checks

If the score rises between rounds without corresponding verifiable improvements (new experiments, fixed citations), flag it as potential reward hacking. Write a warning to RESEARCH.md Context.

Limits

Run at most 4 review rounds per session. After round 4, escalate to the user regardless of score.
After each round, write the score and top weaknesses to RESEARCH.md Context.

Example

Input: paper.md with 3 quantitative claims; 1 claim lacks a matching RESEARCH.md result. Output: score=6/10, verifiable_failures=["Table 2 improvement not in Context"], recommended_action=REFINE.

---name: co-reviewdescription: Adversarial critique of a research draft or paper using an isolated critic context. The critic must have no access to the current conversation history — context isolation, not model identity, is what makes the review independent. Use when a draft is complete or partially complete, after each writing round, or when `co-write` invokes it automatically. Returns FATAL/MAJOR/MINOR issues, a structured score (0–10), and a PROCEED/REFINE/PIVOT recommendation. Caps at 4 rounds. Trigger phrases: "review paper", "critique draft", "check citations", "adversarial review".---# Co-ReviewRun an adversarial review in an **isolated context**. The critic must have no access to the current conversation history — this is what makes the critique independent, not which model is used.## Critic Selection1. Use the first available isolated critic in priority order: - Claude subagent (via `Task` tool or similar) with a clean context — preferred when no external MCP is configured, same model is fine - `codex:codex` (Codex MCP) — isolated by the MCP boundary - `auto-review-loop-llm` skill (any OpenAI-compatible endpoint) - `auto-review-loop-minimax` skill (MiniMax)2. Record which critic was used in `RESEARCH.md` **Context** alongside the score.3. If no isolated context can be established, say so and **stop**. Do not review within the same conversation context.## Review Rubric (Fixed — Do Not Modify)3. Send the work to the critic with this exact rubric:**VERIFIABLE** (graded on evidence, not opinion):- Every claim maps to an experiment result in `RESEARCH.md` **Context**. Flag claims that don't.- No hallucinated citations. Check every DOI and arXiv ID. Flag any that don't resolve.- Ablations support stated conclusions. Check the numbers match.**SUBJECTIVE** (graded by the critic):- Clarity of problem statement (1–5)- Novelty relative to cited work (1–5)- Writing quality (1–5)4. The critic must return a structured result matching the schema in `RUBRIC.md`. See `RUBRIC.md` for field definitions, the PROCEED/REFINE/PIVOT decision rules, and the verbatim integrity instruction to include in every critic prompt.## Integrity Checks5. If the score rises between rounds without corresponding verifiable improvements (new experiments, fixed citations), flag it as **potential reward hacking**. Write a warning to `RESEARCH.md` **Context**.## Limits6. Run at most **4 review rounds** per session. After round 4, escalate to the user regardless of score.7. After each round, write the score and top weaknesses to `RESEARCH.md` **Context**.## ExampleInput: paper.md with 3 quantitative claims; 1 claim lacks a matching RESEARCH.md result.Output: score=6/10, verifiable_failures=["Table 2 improvement not in Context"], recommended_action=REFINE.

Frequently asked questions