What does this skill do, and when should you use it?
SkillCheck Free is an installable skill/rule file that turns your AI assistant into a validator for Claude Code skills. It runs fully offline: it only reads files (Read/Glob/Grep), makes no external API calls, and needs no keys. Validation covers YAML frontmatter spec conformance, description trigger quality, contradictions and vague wording, knowledge density, and prompt-injection risks like hidden Unicode (ASCII smuggling, BiDi overrides, zero-width characters). Output is a per-skill report where each issue carries a check ID, severity, line number and fix, with good practices listed separately as strengths. Deeper features such as anti-slop detection, security scanning and Eval Kit belong to the paid SkillCheck Pro; this repo contains only the Free tier rules.
- Locates target SKILL.md files with Glob (plain and namespaced paths) and loads any references/*.md in the same folder
- Validates frontmatter: name/description format, agentskills spec fields, Claude Code and community extension fields, category and produces/consumes formats
- Runs semantic checks: contradictory instructions, ambiguous terms, hollow wisdom/platitudes, misplaced routing content, restraint without a safety carve-out
- Runs security checks: XML angle brackets in frontmatter, Unicode tag-character smuggling (decodes the hidden payload), Trojan-source BiDi controls, zero-width characters
- Produces one report per skill with check ID, Critical/Warning/Suggestion severity, line number and fix; strengths reported in a separate list
- A skill author running a final offline conformance check on a SKILL.md before publishing, with actionable per-issue fixes
- An engineer maintaining a team skill library validating a whole skills directory in one pass, one report per skill
- A security-minded reviewer using the Cat 30 steganography checks to expose instructions hidden in invisible Unicode characters
- A Cursor developer copying cursor-rules/skill-check.mdc into .cursor/rules to validate skills inside their project
- A Claude Projects user pasting the rules into custom instructions to validate skill definitions directly on claude.ai
- Teams needing anti-slop detection, security scanning, token budget analysis, WCAG compliance or Eval Kit — these are explicitly SkillCheck Pro features, absent from the free tier
- Anyone wanting LinkedIn skill engagement analysis — the description redirects that to a separate skillcheck-engage skill
- CI/CD pipeline users — a standalone CI/CD binary is a Pro feature; the free tier runs only interactively through an AI assistant
How do you install this skill?
- Static review executed nothing; actual check behavior and false-positive rates must be verified by the user. Author-disclosed ~5% false-positive rates exist for heuristic checks (4.6/4.9).
- references/cli.md points to a separate repo (skillcheck-cli) whose relationship and maintenance status are unexplained; do not conflate it with this skill.
- No CHANGELOG; version 3.32.0 has no traceable history; publisher identity is unverified by any registry.
- The skill is English-only; applicability of semantic checks to Chinese-language skills is undeclared.
- Local filesystem
Simplest path, via the skills CLI:
Claude Code
npx skills add olgasafonova/SkillCheck-FreeOr copy the skill folder manually:
Claude Code
git clone https://github.com/olgasafonova/SkillCheck-Free.git
cp -r SkillCheck-Free/skills/skill-check ~/.claude/skills/No file installation needed — paste the content after the frontmatter into Project custom instructions:
Claude.ai (Projects)
# Open claude.ai → Project settings → Custom instructions
# Paste everything after the --- frontmatter in skills/skill-check/SKILL.mdCopy the rule file into your project's rules directory:
Cursor
git clone https://github.com/olgasafonova/SkillCheck-Free.git
mkdir -p .cursor/rules
cp SkillCheck-Free/cursor-rules/skill-check.mdc .cursor/rules/How do you use this skill?
Once installed, send your agent any of these to trigger it:
- skillcheck my skill at ~/.claude/skills/weekly-reports/SKILL.md
- check skill at path/to/SKILL.md and show me all Critical and Warning issues
- validate my skills in ~/.claude/skills/ — one report per skill
- skillcheck this skill folder and list its strengths as well
After installing, say "skillcheck my skill" or "validate my skills" in Claude Code, Claude Desktop/Web, or Cursor. The skill runs a four-step flow: Step 1 locates target SKILL.md files with Glob; Step 2 reads the skill body and any references/ files; Step 3 applies each rule section; Step 4 emits the report. It performs at most one validation pass per skill and stops once every located skill has been reported; unreadable SKILL.md files are noted and aborted in the report. Input can be a path to a single SKILL.md or a directory of skills (one report per skill). The check is read-only and idempotent — safe to re-run.
What are this skill's strengths and limitations?
- Fully offline and read-only: no API keys, env vars or network calls — safe for security-sensitive environments
- Fine-grained, explainable rules: every finding ships with check ID, severity, line number and a concrete fix
- Distinctive coverage: steganographic Unicode / Trojan Source checks most linters don't have
- Cross-host: the same rules adapt to Claude Code, Claude Desktop/Web (Projects) and Cursor
- Idempotent and safe to re-run; reports are identical for unchanged files
- Depth is deliberately reserved for Pro: anti-slop, security scanning, token analysis, Eval Kit and the CI/CD binary are all absent from the free tier
- Results are produced by an AI assistant following a rule file, not a deterministic binary — consistency across models/sessions depends on instruction-following
- The repo documents no standalone test suite or cross-platform verification evidence; install reach relies on third-party registries like skills.sh
- Lightweight resources, but the author must interpret validation reports; newcomers to the agentskills spec may need to read the spec first
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| SkillCheck Free this page | 64 · Recommended | ★ 41 | 14d ago | MIT |
| Skill Spec Validator | 45 · Not recommended | ★ 1.7k | 17d ago | MIT |
| Claude Code Plugin Skill Development ✓ Anthropic · Official | 51 · Use with care | ★ 38k | 11d ago | Apache-2.0 |
| Claude Code Skill Development ✓ Anthropic · Official | 46 · Use with care | ★ 151k | 11d ago | — |
| Claude Skill Template Example ✓ Anthropic · Official | 58 · Recommended | ★ 38k | 11d ago | Apache-2.0 |
This is the free tier of SkillCheck Pro (getskillcheck.com): Pro adds anti-slop detection, security scanning, token budget analysis, WCAG, enterprise/production readiness checks, an Eval Kit, and a CI/CD binary. The free tier keeps structure, semantic, naming and knowledge-density checks (hollow content), the 8 deterministic OWASP Agentic Top 10 checks, and all invisible-Unicode checks.
How did FollowSkills review this skill?
The skill declares read-only tools (Read Glob Grep), no network access, no env vars, no side effects; workflow includes execution limits and idempotency; data flow is explicit. Deducted: unverified publisher, attribution relies on self-declared author metadata only, no rollback/confirmation mechanisms documented (largely unnecessary for read-only checks, but false-positive handling for security checks rests on heuristic claims). No red-line risks found.
SKILL.md is internally consistent: four-step workflow, execution caps, error table, failure feedback (report when a file cannot be read), and disclosed false-positive rates (4.6, ~5%). Deducted: static review only, no committed test suite covering key check paths, correctness of regex rules (Unicode ranges, BiDi exemptions) unverifiable, and heuristic checks depend on model execution which introduces nondeterminism.
Triggers are explicit (check skill, skillcheck, validate SKILL.md), negative triggers clearly exclude Pro/LinkedIn scopes, input/output contract defined, fully offline so mainland-China reachability is not affected. Deducted: English only, no Chinese-language support; boundaries around Claude Code extension fields are documented but utility is narrower on other agents.
MIT license, version 3.32.0, author attribution, maintenance path (GitHub issues), layered docs (SKILL.md + references files), and Free/Pro scope disclosure are all present. Deducted: no CHANGELOG, the CLI referenced in references/cli.md links to a separate repo whose relationship is unexplained, no version history backing the 3.32.0 claim, minor rule overlap between marketplace-checks.md and the main file.
The core task (validating SKILL.md against the agentskills spec) has concrete rules, worked examples, and a defined output format; marginal value over manual review is clear. Deducted: static review cannot verify representative output correctness/completeness, the 0-100 summary score has no defined calculation rule, and output still requires human review.
Rules ship with a worked-examples file, README sample output, and registry inclusion claims. Deducted: no CI workflow tests the checks themselves (the present CI is only a Claude code-review bot, not execution evidence), no independent third-party reproduction material, and the registry listings are not corroborated within the repository.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →