Set up
Skills are installed by placing their folder where your agent loads skills.
The download comes from github.com/carlosmarte/agents-skills-kit itself, not from our server, so what you get is exactly what its author published.
Install instructions come from the README
README
synced Oct 23, 2026agents-skills-kit
Polyglot toolkit for authoring, validating, and distributing Agent Skills — the SKILL.md-based format described in agentskills.io.md. Each skill lives under .agents/skills/<name>/ and is mirrored into .claude/skills/<name> as a relative symlink so Claude Code's harness auto-discovers it. The runtime CLI (skills-ref) ships parity-tested mjs and py implementations under agent-skill-kit-commons.
Skills
| Skill | What it does |
|---|---|
agent-skill-kit-ci-integrator | Wire the agent-skills toolkit into CI — install the GitHub Actions workflow that runs validate/lint/audit/deps on every PR, and install the pre-commit hooks for local-loop parity. |
agent-skill-kit-commons | Shared assets and reference implementations consumed by every other skill in the agent-skills-kit. |
agent-skill-kit-creator | Scaffold and validate agent skills. Tier/scope (company, enterprise, org, team, project, application) is optional and off by default — opt in with --tiers / --agents-skills-kit-tiers. |
agent-skill-kit-dependency-resolver | Resolve and topologically sort skill dependencies, enforce the spec §3.3 inversion rule (project→app→team→org), warn on unpinned versions, and fail-closed on incompatible range conflicts. |
agent-skill-kit-description-optimizer | Optimize a SKILL.md description for trigger quality — generate synthetic positive and negative queries, score the current description by hit-rate vs false-positive, and propose three ranked rewrites. |
agent-skill-kit-discoverability-linter | Score a SKILL.md description 0-100 for trigger quality, flag vague language, warn on vendor-specific frontmatter, and verify relative file references resolve. |
agent-skill-kit-prompt-emitter | Emit progressive-disclosure artifacts from a skills tree — an <available_skills> XML block for stage-1 agent startup, or canonical JSON of frontmatter properties for CI consumption. |
agent-skill-kit-security-auditor | Audit a skills tree for the four canonical security findings — unauthorized frontmatter fields, command-injection patterns in scripts, unvetted cross-repo dependency origins, and host-path leaks in body or scripts. |
agent-skill-kit-validator | Validate one or many SKILL.md files against the agentskills.io specification — frontmatter, naming, body, optional-directory, and reference checks. |
Install
Per skill — npx skills add
Install any single skill into Claude Code:
npx skills add carlosmarte/agents-skills-kit \
--skill agent-skill-kit-ci-integrator -a claude-code \
--skill agent-skill-kit-commons -a claude-code \
--skill agent-skill-kit-creator -a claude-code \
--skill agent-skill-kit-dependency-resolver -a claude-code \
--skill agent-skill-kit-description-optimizer -a claude-code \
--skill agent-skill-kit-discoverability-linter -a claude-code \
--skill agent-skill-kit-prompt-emitter -a claude-code \
--skill agent-skill-kit-security-auditor -a claude-code \
--skill agent-skill-kit-validator
Swap --skill for any name from the table above.
One-shot install via curl
curl -fsSL https://raw.githubusercontent.com/carlosmarte/agents-skills-kit/main/install.sh | bash
Preview only — change nothing:
curl -fsSL https://raw.githubusercontent.com/carlosmarte/agents-skills-kit/main/install.sh | bash -s -- --dry-run
Layout
.agents/skills/<name>/SKILL.md # source of truth for each skill
.claude/skills/<name> -> ../../.agents/skills/<name> # relative symlink (harness-discovered)