agentleFS
yeisonrestrepoyeisonrestrepo/code-conductor/skills/critical-review/SKILL.md

critical-review

Skill17 starsChanged 37 days agoMITView on GitHub

4-phase adversarial review protocol: pre-flight analysis, adversarial execution check, self-correction loop, and [VALIDATION] report

5 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

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

  • 0

    Common

    Most of the 5 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Before writing any code, perform a Vulnerability Assessment on the proposed solution.

    line 15
  2. Distinctive:

    After completing the initial implementation, switch to an Adversarial Review persona.

    line 29
  3. Distinctive:

    Do not bundle multiple fixes into one step.

    line 60
  4. Very specific:

    Do not omit [VALIDATION]. If verbosity is MIN, keep each item to one line.

    line 75

“Common” is measured: kept by at least 6 in 10 of the 5 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

Critical Review Protocol

Apply this protocol to every implementation task. Do not skip phases.


Phase 1 — Pre-Flight Analysis

Before writing any code, perform a Vulnerability Assessment on the proposed solution. Do not proceed to implementation until all three are answered:

Happy Path — Describe the most common successful execution flow in one sentence.

Failure Points — Where is this logic most likely to break? List: race conditions, null/undefined references, network timeouts, missing auth, unhandled promise rejections, type coercion surprises.

Boundary Conditions — What inputs or volume break assumptions? List: empty arrays, zero values, extremely large payloads, concurrent calls, missing environment variables.

Present the Pre-Flight Analysis and wait for confirmation before implementing.


Phase 2 — Execution & Adversarial Review

After completing the initial implementation, switch to an Adversarial Review persona. Run each check:

RESILIENCE

Does the system fail gracefully or crash silently when a dependency fails or input is malformed? Look for:

  • Caught errors that swallow the stack trace
  • Unchecked nulls after async calls
  • Missing fallbacks on external API responses

EFFICIENCY

Is this the best possible outcome? Flag:

  • Redundant loops or duplicate traversals
  • Deep nesting (>3 levels) replaceable by early returns
  • High cyclomatic complexity (branches > 5 in one function)
  • N+1 query patterns

FRICTION

Analyze the Happy Path for unnecessary friction:

  • Steps the caller must repeat on every use
  • Configuration that should have a sensible default
  • Error messages that don't tell the user what to do next

Phase 3 — Self-Correction Loop

For each weakness identified in Phase 2:

  1. State the weakness in one sentence (file:line reference if applicable).
  2. Refactor the code to address it.
  3. Re-verify: confirm the original failure point no longer applies.

Do not bundle multiple fixes into one step. One weakness → one refactor → one verification.


Phase 4 — Final Report

End every implementation with a [VALIDATION] section:

[VALIDATION]
- Edge cases accounted for: [list]
- Why this is the best outcome vs simpler alternatives: [one paragraph]
- Residual risks outside current scope: [list or "none identified"]

Do not omit [VALIDATION]. If verbosity is MIN, keep each item to one line. If VERBOSE, expand freely.

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.