The empirical chain — every output has a URL, the on-ramp doesn't Every artifact has a URL. The path to authoring them doesn't. A new contributor must read 5+ blog posts + the spec + agent-private memory files to assemble "how to add a feature". ARTIFACT WHERE IT LIVES CANONICAL DOC? SAMA spec (rules + profile + verifier) /sama/v2 ✓ on-site §4 verifier verdict (live proof) /sama/v2/verify ✓ live · auditable §5 metric definitions + measurements /sama/v2 §5 · /blog ✓ on-site /goal contracts (post PR #47) /goals · /goals/<slug> ✓ archived Source code (every file) /GIT/tdd.md/tree/main ✓ browseable Discoverability (every URL) /sitemap.xml ✓ generated Blog narrative (every decision) /blog ✓ indexed Authoring workflow (the /goal rule) agent-private memory file ~ not on site Layer / test / branch / deploy patterns scattered across PRs + posts ~ no canonical doc "How do I contribute?" — the on-ramp nowhere — assemble from 11+ sources ✗ MISSING The self-violation: "Every artifact is auditable" — yes, every OUTPUT is. The path to producing the outputs requires reverse-engineering across blog posts + memory files (which humans can't even read) + commit history. A new contributor lands on tdd.md with six entry points and no front door. The chain has all its links — except the one that brings a contributor in. https://tdd.md