Docforge Api Docs

Quick Reference

1. Discovery (CRITICAL)

• discovery-three-tier - Implement the three-tier discovery pattern (llms.txt, llms-full.txt, docs.md)
• discovery-well-known - Serve .well-known/ endpoints for machine discovery
• discovery-link-header - Include Link headers pointing to llms.txt

2. Route Scanning (HIGH)

• scanning-hono-routes - Extract endpoints from Hono route registrations
• scanning-middleware-context - Capture auth, rate limits, and middleware from route chains
• scanning-request-response - Infer request/response schemas from handler code

3. Documentation Output (HIGH)

• output-llms-txt - Format llms.txt as a concise endpoint directory
• output-llms-full - Format llms-full.txt with complete request/response details
• output-json-schema - Generate JSON Schema for request/response bodies
• output-curl-examples - Include working curl examples for every endpoint

4. Quality Standards (MEDIUM)

• quality-no-truncation - Never truncate response examples or field lists
• quality-real-values - Use realistic example values, not placeholders
• quality-error-responses - Document all error codes and response shapes
• quality-auth-documented - Always document authentication requirements

5. Automation (MEDIUM)

• automation-git-hook - Generate docs on commit via pre-push hook
• automation-cron-refresh - Schedule periodic doc regeneration
• automation-diff-detection - Only regenerate docs when routes change

How to Use

Read individual rule files for detailed explanations and code examples:

rules/discovery-three-tier.md
rules/scanning-hono-routes.md
rules/output-llms-txt.md

Each rule file contains:

1. Brief explanation of the rule
2. Why it matters for agent consumption
3. Incorrect example (what to avoid)
4. Correct example (what to do)
5. Additional context and references