# Agent Exchange Skills

> Three Agent Skills (SKILL.md, agentskills.io format) for using Agent Exchange services from Claude Code, Codex, Cursor, OpenClaw or any agent. Served from this domain as a git repository (smart HTTP v2, dumb-HTTP fallback), a Claude Code plugin marketplace, zip archives and plain files. No registry, no review queue, no GitHub. Version 1.0.0, commit fbce5b3b136e.

## Skills
- [agent-readiness-grade](https://skills.agentexchange.work/skills/agent-readiness-grade/SKILL.md): Grade any public website A-F (0-12) on how well AI crawlers and agents can read, cite and transact with it, using grade.agentexchange.work. Checks robots.txt access for AI bots, machine-readable surfaces (llms.txt, sitemap.xml, openapi.json, .well-known/x402, agent-card.json, mcp-registry-auth), readability without JavaScript, structured data (JSON-LD, Open Graph), published legal notices and MCP census presence. Use when asked whether a site is agent-ready or AI-readable, to audit llms.txt or robots.txt for AI crawlers, to explain a grade, or to compare sites.
- [agentexchange-web-tools](https://skills.agentexchange.work/skills/agentexchange-web-tools/SKILL.md): Read JavaScript-rendered web pages, take screenshots and extract structured data through a real browser at tools.agentexchange.work, paying per call with USDC on Base (x402). No API key, no account. Use when a page needs JavaScript to render, when a screenshot is needed, or when structured data should be pulled from a page without writing a parser.
- [compliance-desk](https://skills.agentexchange.work/skills/compliance-desk/SKILL.md): Build, send and track legally-deadlined requests through takeitdown.agentexchange.work (Compliance Desk) - TAKE IT DOWN Act removal of intimate images posted without consent (48-hour deadline), DMCA takedown notices, CCPA/GDPR privacy requests (access, deletion, correction, opt-out of sale) and TCPA/CAN-SPAM do-not-call or unsubscribe requests. Also checks whether a website publishes the notices and request mechanisms it is legally required to have, and sets up a hosted, logged request desk for a site the user operates. Use when a user wants content removed, a takedown or privacy request written, unwanted calls, texts or emails stopped, a site's compliance notices checked, or a takedown, DSA, privacy, accessibility or opt-out intake for their own site.

## Install
- Any git client: git clone https://skills.agentexchange.work/agentexchange-skills.git
- Claude Code: claude plugin marketplace add https://skills.agentexchange.work/agentexchange-skills.git ; then claude plugin install <skill>@agentexchange. Or add https://skills.agentexchange.work/marketplace.json (zip archives, no git).
- OpenClaw: openclaw skills install git:https://skills.agentexchange.work/<skill>.git (one repository per skill, SKILL.md at its root; append @main or @v1.0.0 to pin), or clone and install ./agentexchange-skills/skills/<skill>.
- Codex, Cursor, OpenClaw (shared directory): clone, then cp -r agentexchange-skills/skills/. ~/.agents/skills/
- Plain HTTP: https://skills.agentexchange.work/skills/<skill>/SKILL.md

## Machine-readable
- [index.json](https://skills.agentexchange.work/index.json): every file with sha256, install commands per client, which are verified by test and which are documented only.
- [marketplace.json](https://skills.agentexchange.work/marketplace.json): Claude Code marketplace (hosted variant, archive sources).
- [.well-known/agentexchange-skills.json](https://skills.agentexchange.work/.well-known/agentexchange-skills.json): discovery document.
- [mcp.json](https://skills.agentexchange.work/mcp.json): our public MCP endpoints (index, measured data, store, tooldrift).

## Services the skills describe
- https://tools.agentexchange.work (x402 web tools) · https://takeitdown.agentexchange.work (Compliance Desk) · https://grade.agentexchange.work (Agent-Readiness Grade)
- Hub of everything we run: https://network.agentexchange.work