thedaviddias/skill-check

When your agent uses it

What it loads into context

Token counts are estimates. Scripts run separately; they are not counted as reference context.

# split-into-references

## Use when

Use when `skill-check` reports `body.max_lines` and you need to modularize a long `SKILL.md` into maintainable `references/*.md` docs after running `split-body`.

## Workflow

1. Run preview first:
   - `npx skill-check split-body <skill-dir-or-file>`
2. Apply deterministic split:
   - `npx skill-check split-body <skill-dir-or-file> --write`
3. Edit generated references docs to improve readability:
   - tighten headings so each file has one clear purpose
   - add a 1-3 line summary at the top of each references file

bad:slug

When your agent uses it

Use when testing invalid slug format validation.

What it loads into context

Token counts are estimates. Scripts run separately; they are not counted as reference context.

# Bad Slug

my-skill

When your agent uses it

Use when testing broken reference links.

What it loads into context

Token counts are estimates. Scripts run separately; they are not counted as reference context.

# Broken Link

See [missing](references/missing.md).

other-skill

When your agent uses it

Use when testing mismatched name validation.

What it loads into context

Token counts are estimates. Scripts run separately; they are not counted as reference context.

# Mismatch

When your agent uses it

What it loads into context

Token counts are estimates. Scripts run separately; they are not counted as reference context.

# Missing Frontmatter

This fixture intentionally has no frontmatter block.

When your agent uses it

What it loads into context

Token counts are estimates. Scripts run separately; they are not counted as reference context.

# Fake Bad Skill

This file intentionally omits frontmatter to trigger validation errors.

FakeBadSlug

When your agent uses it

Use when testing failures for bad slug formatting and mismatched names in intentionally broken fixtures.

What it loads into context

Token counts are estimates. Scripts run separately; they are not counted as reference context.

# Fake Bad Slug

This fixture intentionally violates the expected slug naming format.

real-review-skill

When your agent uses it

Use when reviewing pull requests for regressions, missing tests, and release-blocking issues across multiple service boundaries.

What it loads into context

Token counts are estimates. Scripts run separately; they are not counted as reference context.

# Real Review Skill

This valid skill exists alongside intentionally invalid fixtures.

my-skill

When your agent uses it

Use when testing body line limits.

What it loads into context

Token counts are estimates. Scripts run separately; they are not counted as reference context.

# Long Body
Line 1
Line 2
Line 3
Line 4
Line 5
Line 6
Line 7
Line 8
Line 9
Line 10
Line 11
Line 12
Line 13

my-skill

When your agent uses it

Use when aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa

What it loads into context

Token counts are estimates. Scripts run separately; they are not counted as reference context.

# Too Long Description

my-skill

When your agent uses it

Use when testing trailing newline validation.

What it loads into context

Token counts are estimates. Scripts run separately; they are not counted as reference context.

# Trailing

good-skill

When your agent uses it

Use when you need a valid fixture skill for passing validation tests.

What it loads into context

Token counts are estimates. Scripts run separately; they are not counted as reference context.

# Good Skill

See [reference](references/guide.md).

agent-scan-guard

When your agent uses it

Use when you need to run security checks for MCP and agent workflows before merging changes in shared repositories.

What it loads into context

Token counts are estimates. Scripts run separately; they are not counted as reference context.

# Agent Scan Guard

Run `skill-check check . --agent-scan --agent-scan-runner npx` before release.

See [playbook](references/playbook.md).

content-qa

When your agent uses it

Use when reviewing long-form marketing or technical content for clarity, consistency, and executable examples before publishing.

What it loads into context

Token counts are estimates. Scripts run separately; they are not counted as reference context.

# Content QA

Check links, command snippets, and section ordering before publishing.

fake-demo-skill

When your agent uses it

Use when you want a clearly fake but valid skill fixture for local smoke demos and parser coverage in tests.

What it loads into context

Token counts are estimates. Scripts run separately; they are not counted as reference context.

# Fake Demo Skill

This fixture is intentionally fictional but follows the expected skill format.

skill-check

When your agent uses it

Use when the user wants to validate, lint, or audit agent skill files (SKILL.md). Use when they say "validate these skills," "check this repo's skills," "lint SKILL.md files," or "audit skills in [repo URL]." Run skill-check locally or against a cloned GitHub repo and summarize findings.

What it loads into context

Token counts are estimates. Scripts run separately; they are not counted as reference context.

# skill-check

Validate agent skill files (SKILL.md) using the skill-check tool. Support both local paths and GitHub repos.

## Installation

Ensure skill-check is available before running checks:

- **No install (recommended for one-off or agent use):** `npx skill-check` — uses npm to run the latest version; requires Node.js and network on first run.
- **Global install (curl):** `curl -fsSL https://raw.githubusercontent.com/thedaviddias/skill-check/main/scripts/install.sh | bash` — installs a global `skill-check` binary.
- **Homebrew:** `brew tap thedaviddias/skill-check https://github.com/thedaviddias/skill-check` then `brew install skill-check`.

If the user has not installed skill-check, use `npx skill-check` so no prior install is needed. To confirm the tool is available, run `npx skill-check rules` and expect a list of built-in rule IDs.