agentleFS
arthur-plautzarthur-plautz/research-toolkit/.claude/skills/critique/SKILL.md

critique

Skill0 starsChanged 4 months agoMITView on GitHub

Adversarially reviews a project document to surface shallow, fragile, or wrong statements before a real reviewer does. Spawns a read-only opus subagent that verifies load-bearing claims against ground truth (code, installed libraries, cited papers) instead of trusting the document. Handles proposals (docs/proposals/) and literature reviews (docs/reviews/) with different rubrics. Invoke with `/critique`, or when the user asks to critique, stress-test, or find weaknesses in a proposal or review.

79 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use delight, teach-impeccable and 10 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.
  • HK-hubSkill

    delight

    Add moments of joy, personality, and unexpected touches that make interfaces memorable and enjoyable to use. Elevates functional to delightful. Use when the user asks to add polish, personality, animations, micro-interactions, delight, or make an interface feel fun or memorable.

    Also in 32 projects that use this

  • HK-hubSkill

    teach-impeccable

    One-time setup that gathers design context for your project and saves it to your AI config file. Run once to establish persistent design guidelines.

    Also in 18 projects that use this

  • zwy123zwySkill

    normalize

    将用户随意输入的提示词规范化为结构清晰、便于大模型执行的标准化提示词。适用于需要整理、统一提示词格式时调用。

    Also in 18 projects that use this

  • owshq-mecSkill

    layout

    Improve layout, spacing, and visual rhythm. Fixes monotonous grids, inconsistent spacing, and weak visual hierarchy. Use when the user mentions layout feeling off, spacing issues, visual hierarchy, crowded UI, alignment problems, or wanting better composition.

    Also in 8 projects that use this

  • +8 more from HK-hub

The file

We split this skill into the rules it gives an agent

3 rules. 0 are common, 1 is distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 72 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Before finalizing a proposal or literature review

    line 22
  2. Very specific:

    Not for: interviewing the user about an unwritten plan (that is /grilling), or for applying fixes (that is the edit mode of /proposals or …

    line 24
  3. Very specific:

    Stop there: Do not edit the document. If the user wants findings applied, point them at /proposals edit <id> or /literature-review edit …

    line 50

“Common” is measured: kept by at least 6 in 10 of the 72 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

Purpose

Get a document torn apart by a hostile-but-fair expert reviewer while it can still be fixed. The value is not opinion — it is verification: load-bearing technical and grounding claims are checked against ground truth (the actual code, the installed library source, the cited paper), not accepted on the document's word. That is what catches the confidently-stated premise that turns out false.

This skill orchestrates; the document-critic opus subagent does the review. The skill never edits the document — critique only. To act on the findings, the user routes to /proposals edit or /literature-review edit.

When to invoke

  • User asks to critique, review, stress-test, or find weaknesses in a document
  • User invokes /critique or /critique <path-or-id>
  • Before finalizing a proposal or literature review

Not for: interviewing the user about an unwritten plan (that is /grilling), or for applying fixes (that is the edit mode of /proposals or /literature-review).

Flow

  1. Resolve the target. Take the path or ID from the user's argument. If none was given, use the document in recent conversation context; if still ambiguous, Glob docs/proposals/*.md and docs/reviews/*.md and ask which one.
  2. Detect the type from the path:
    • under docs/proposals/ → proposal
    • under docs/reviews/ → literature-review
    • anything else → generic
  3. Spawn the document-critic agent (subagent_type document-critic), passing the absolute document path and the detected type. Run it and wait for its result — the user is waiting on the critique, so do not proceed without it. The agent is read-only; it reads the doc, the project docs it stands on, the code / installed libraries its technical claims depend on, and the cited papers, then reports findings ordered by severity.
  4. Relay the findings to the user. The agent's report is not shown to the user directly — surface it: lead with the single highest-priority fix, then the severity-ordered findings with their evidence, then what held up. Do not water it down and do not pad it.
  5. Stop there. Do not edit the document. If the user wants findings applied, point them at /proposals edit <id> or /literature-review edit <id>.

Notes

  • The rubrics (proposal vs. literature-review vs. generic) live in the document-critic agent definition, keyed by the type passed in step 3.
  • Spawn one agent per document. For several documents, spawn several — do not ask one agent to review a batch, so each critique stays focused and its verification deep.

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.