generated: '2026-09-01'
method: generated
source: >-
  Generated by API Evangelist from the provider's own OpenAPI 3.1 contract
  (https://www.pathwren.workers.dev/openapi.json), MCP tools/list responses and
  /.well-known/api-onboarding, all fetched 2026-09-01. The provider publishes no skills/ directory
  and no AGENTS.md — /AGENTS.md, /agents.md and /skills.json each returned 404 on 2026-09-01 — so
  these are generated, not harvested. Every operationId named in every skill appears in the
  provider's contract; none were invented.
api: AI Crawler Index
api_base: https://www.pathwren.workers.dev
skill_count: 4
skills:
- name: identify-crawler
  file: skills/pathwren-identify-crawler.md
  operations: [listCrawlers, getCrawler, getUserAgentRegex]
  mirrors_mcp_tool: classify_user_agent
- name: build-robots-policy
  file: skills/pathwren-build-robots-policy.md
  operations: [getPolicy, getRobotsPolicy, listCrawlers]
  mirrors_mcp_tool: generate_robots_txt
- name: verify-crawler-ip
  file: skills/pathwren-verify-crawler-ip.md
  operations: [getAllIpRanges, getAllIpRangesText, getIpRangeSource, listIpSources, getCrawler]
  mirrors_mcp_tool: is_verified_crawler_ip
- name: monitor-index-freshness
  file: skills/pathwren-monitor-index-freshness.md
  operations: [getStatus, getFeed, listIpSources]
  mirrors_mcp_tool: whats_changed
note: >-
  Each skill mirrors one of the six tools the provider's own MCP server exposes, expressed against
  the REST contract for agents that do not speak MCP. The two remaining tools — lookup_crawler and
  list_crawlers — are single-operation calls folded into identify-crawler rather than given their
  own skill. See mcp/pathwren-tool-crosswalk.yml for the full binding.