Passlock skill for Codex, Claude and other LLM coding agents

Install with skills.sh (option 1)

To install the skill into your agent’s skills directory so it loads automatically, use skills.sh:

npm

npx skills add passlock-dev/passlock --skill passlock

yarn

yarn dlx skills add passlock-dev/passlock --skill passlock

pnpm

pnpx skills add passlock-dev/passlock --skill passlock

bun

bunx skills add passlock-dev/passlock --skill passlock

This works with Claude Code, Cursor, Codex, GitHub Copilot, Windsurf, Gemini, and other agents that support the Agent Skills specification.

Download the skill (option 2)

Use our SKILL.md to instruct your LLM coding agent to read the authoritative, agent-friendly documentation.

Where to place the skill

Copy this skill into the relevant location:

• Codex project: .agents/skills/passlock/SKILL.md
• Codex user: ~/.agents/skills/passlock/SKILL.md
• Claude Code project: .claude/skills/passlock/SKILL.md
• Claude Code user: ~/.claude/skills/passlock/SKILL.md

How the skill works

The skill is a public-doc retrieval guide for Passlock. It tells an agent to:

• Use the public Passlock docs and API reference markdown files
• Start with https://passlock.dev/llms.txt for broad discovery
• Search the generated LLM indexes on passlock.dev and apidocs.passlock.dev
• Fetch only the relevant markdown pages after search narrows the topic
• Prefer runbook docs for integration guidance and API docs for exact symbols and signatures
• Cite public URLs in the final answer

Local search index

The skill maintains a local search index, so agents can quickly identify the relevant documentation. It does not cache the documents themselves. At query time the agent:

• Fetches llms/search-cache-id.txt
• If search-index-{id}.json does not exist locally:
  • Fetches search-index-{id}.json
  • Caches that file locally under the same ID
• Searches the cache for query terms
• Returns the highest-scoring matching page URLs

Note: The content of search-cache-id.txt serves as a cache busting key