All skills
richtabor avatar

/skill-review

@787e3d5
by Rich Taborrichtabor/agent-skills72 stars
10

Reviews and validates agent skills against best practices. Triggers on "review this skill", "check my skill", "validate skill", "is this skill well-written", or when creating/editing skills.

Use this Skill: https://skilld.dev/gh/richtabor/agent-skills/skill-review

This session only. Nothing lands on disk.

referenceschecklist.md

≈706 tokens on demand. Your agent reads this file only when SKILL.md points to it.

Skill Review Checklist

Quick validation checklist for reviewing skills. Use for fast pass/fail audits.


Structure

□ SKILL.md exists
□ Directory name matches frontmatter `name`
□ No extraneous files (README.md, CHANGELOG.md, etc.)
□ References one level deep (no chains)
□ Forward slashes only (no Windows paths)
□ Script paths referenced in body exist in directory
□ External binary dependencies documented

Frontmatter

□ name: lowercase with hyphens only [a-z0-9-]
□ name: ≤64 characters
□ name: no "anthropic" or "claude"
□ description: non-empty
□ description: ≤1024 characters
□ description: third-person voice
□ description: includes what it does
□ description: includes when to trigger

Description Quality

□ Specific actions (not vague like "helps with")
□ Trigger phrases included ("Use when...")
□ Synonyms for discoverability
□ No first/second person ("I can", "You can")

Body Content

□ Under 500 lines total
□ Quick start near top
□ Details in references/ (not front-loaded)
□ Long refs (>100 lines) have TOC
□ No obvious explanations Claude knows
□ Examples over prose
□ Defaults provided when multiple options

Anti-Patterns (should NOT find)

□ No Windows paths (scripts\file.py)
□ No nested reference chains
□ No time-sensitive statements ("before August 2025")
□ No magic numbers without explanation
□ No "you can use X or Y or Z" without default
□ No inconsistent terminology
□ No user-facing docs in skill directory

Optional Elements

□ scripts/ - executable code (if needed)
□ references/ - loaded documentation (if needed)
□ assets/ - output templates (if needed)
□ allowed-tools scoped appropriately (if used)

Quick Scoring

Category Weight Pass/Fail
Frontmatter valid Required
Description quality Required
Body <500 lines Required
No anti-patterns Required
Progressive disclosure Recommended
Token efficiency Recommended

Pass: All required + majority of recommended Needs work: Any required failing Fail: Multiple required failing


Description Template

description: >-
  [What it does - specific actions and capabilities].
  Use when [trigger phrases, contexts, file types, user intents].

Example Passing Description

description: >-
  Reviews UI for accessibility issues against WCAG 2.1/2.2 AA.
  Triggers on "is this accessible?", "check accessibility",
  or contrast/a11y review requests.

Example Failing Description

# Too vague
description: Helps with documents

# Missing triggers
description: Processes PDF files

# Wrong voice
description: I can help you analyze data

Source: SKILL.md on GitHub

1 warning9d4 checks · Risk SAFE
  • Gen Agent Trust Hub9d

    The skill is a comprehensive toolkit for auditing and improving other AI agent skills. It provides structured workflows, checklists, and best practices. No malicious patterns, data exfiltration, or dangerous execution triggers were detected.

  • Socket9d

    No alerts

  • Snyk9d

    Risk: LOW · No issues

  • Runlayer7mo

    4/4 files flagged

Signed by skilld at 787e3d5. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 2 months ago.

Steadyupdated 8 months ago

README badge

README badge for richtabor/agent-skills/skill-review