skill-check
Validate Claude Code skills against the agentskills specification. Catches structural, semantic, and naming issues before users do.
- License 许可
- MIT
- 最近更新
- 2026-11-10
- Stars
- ★ 1
- Forks
- 0
01 前言 Front Matter
- name
- skill-check
- risk
- safe
- tags
- [ "validation", "linter", "agentskills", "skill-authoring", "code-quality" ]
- tools
- [ "claude", "cursor", "windsurf", "codex-cli" ]
- author
- olgasafonova
- source
- https://github.com/olgasafonova/SkillCheck-Free
- license
- MIT
- category
- development
- date_added
- 2026-03-11
- description
- Validate Claude Code skills against the agentskills specification. Catches structural, semantic, and naming issues before users do.
- allowed-tools
- Read Glob
- compatibility
- claude-code
03 SKILL.md 正文
SkillCheck
Overview
Validate SKILL.md files against the agentskills specification and Anthropic best practices. Catches structural errors, semantic contradictions, naming anti-patterns, and quality gaps in a single read-only pass.
When to Use This Skill
- Use when user says "check skill", "skillcheck", or "validate SKILL.md"
- Use when reviewing a skill before publishing to a marketplace
- Use when debugging why a skill doesn't trigger correctly
- Use when onboarding a team to skill authoring standards
- Do NOT use for anti-slop detection, security scanning, or token analysis; use SkillCheck Pro for those
How It Works
Step 1: Parse
Read the target SKILL.md file and extract YAML frontmatter.
Step 2: Validate
Apply all Free tier checks in order:
| Category | Checks | What it catches |
|---|---|---|
| Structure (1.x) | Name format, description WHAT+WHEN, allowed-tools, categories, XML injection | Malformed frontmatter, missing fields |
| Body (2.x) | Line count, hardcoded paths, stale dates, empty sections, deprecated syntax, MCP tool qualification | Content quality issues |
| Naming (3.x) | Vague terms, single-word names, gerund suggestions | Poor discoverability |
| Semantic (4.x) | Contradictions, ambiguous terms, missing output format, wisdom/platitudes, misplaced triggers | Logical inconsistencies |
| Quality (8.x) | Examples, error handling, triggers, output format, prerequisites, negative triggers | Strengths (positive patterns) |
Step 3: Score
Calculate overall score (0-100). Penalties: critical = -20, warning = -5, suggestion = -1.
Step 4: Report
Return structured results: score, grade (Excellent/Good/Needs Work/Poor), issue list with check IDs, line numbers, messages, and fix suggestions.
Examples
Example 1: Validating a skill
User: check my skill at ~/.claude/skills/weekly-report/SKILL.md
SkillCheck output:
## weekly-report Check Results [FREE]
Score: 85/100 (Good)
### Warnings (2)
- 1.2-desc-when (line 3): Description missing WHEN clause
- 4.5-desc-no-triggers (line 3): Description lacks triggering conditions
### Suggestions (1)
- 3.4-gerund-naming (line 2): Skill name could use gerund form
### Passed Checks: 28
Example 2: Clean skill passes all checks
User: skillcheck ~/.claude/skills/processing-pdfs/SKILL.md
Score: 100/100 (Excellent)
All 31 checks passed. No issues found.
Limitations
- Read-only: does not modify any files
- Free tier covers structural, semantic, and naming checks only
- Anti-slop, security, WCAG, token, enterprise, and workflow checks require SkillCheck Pro
- Semantic checks (contradiction detection, wisdom/platitude) are heuristic with ~5% false positive rate
- Does not validate referenced files or scripts; only checks SKILL.md content
- Single-file validation; does not cross-check against other skills in the same directory
Best Practices
- Run SkillCheck before submitting skills to any marketplace
- Fix all critical and warning issues; suggestions are optional
- Use the check ID (e.g.,
1.2-desc-when) to find the exact rule in the skill body - Re-run after fixes to confirm the score improved
Common Pitfalls
-
Problem: Score seems low due to many suggestions Solution: Suggestions cap at -15 points total. Focus on warnings and criticals first.
-
Problem: False positive on ambiguous terms inside code blocks Solution: SkillCheck skips code blocks and inline code. If you still see false positives, wrap the term in backticks.
-
Problem: Wisdom/platitude check flags legitimate instructions Solution: Rephrase generic advice ("Remember that testing is important") as concrete directives ("Run tests before committing").
04 同仓库更多 Skills
vercel-sandbox
Run agent-browser + Chrome inside Vercel Sandbox microVMs for browser automation from any Vercel-deployed app. Use when the user needs browser automation in a Vercel app (Next.js, SvelteKit, Nuxt, Remix, Astro, etc.), wants to run headless Chrome without binary size limits, needs persistent browser sessions across commands, or wants ephemeral isolated browser environments. Triggers include "Vercel Sandbox browser", "microVM Chrome", "agent-browser in sandbox", "browser automation on Vercel", or any task requiring Chrome in a Vercel Sandbox.
slack
Interact with Slack workspaces using browser automation. Use when the user needs to check unread channels, navigate Slack, send messages, extract data, find information, search conversations, or automate any Slack task. Triggers include "check my Slack", "what channels have unreads", "send a message to", "search Slack for", "extract from Slack", "find who said", or any task requiring programmatic Slack interaction.
electron
Automate Electron desktop apps (VS Code, Slack, Discord, Figma, Notion, Spotify, etc.) using agent-browser via Chrome DevTools Protocol. Use when the user needs to interact with an Electron app, automate a desktop app, connect to a running app, control a native app, or test an Electron application. Triggers include "automate Slack app", "control VS Code", "interact with Discord app", "test this Electron app", "connect to desktop app", or any task requiring automation of a native Electron application.
dogfood
Systematically explore and test a web application to find bugs, UX issues, and other problems. Use when asked to "dogfood", "QA", "exploratory test", "find issues", "bug hunt", "test this app/site/platform", or review the quality of a web application. Produces a structured report with full reproduction evidence -- step-by-step screenshots, repro videos, and detailed repro steps for every issue -- so findings can be handed directly to the responsible teams.
agent-browser
Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction.
shadcn
Manages shadcn components and projects — adding, searching, fixing, debugging, styling, and composing UI, including chat interfaces. Provides project context, component docs, and usage examples. Applies when working with shadcn/ui, component registries, presets, --preset codes, or any project with a components.json file. Also triggers for "shadcn init", "create an app with --preset", "switch to --preset", and for building a design system or applying a DESIGN.md or brand spec to shadcn/ui.