docs(self-eval): label terminal/repair/eval checks as documentation-completeness

[SollanSystems]

Closes #35.

Scope-honesty relabel of the three self_eval.py checks that are documentation-completeness scans, with zero behavioral change (the pass/fail logic of every check is byte-unchanged):

• Check labels renamed to say what they prove: terminal-states-documented, repair-record-fields-documented, eval-layers-and-metrics-documented (still exactly 13 checks, labels unique; check_license_present / check_readme_differentiation untouched — test_self_eval calls them by name).
• A comment above each of the three functions states the scope: canonical vocabulary present in the skill prose, not behavioral/runtime enforcement. check_repair_fields' comment also notes its second half (the examples/ failure-mode taxonomy scan) is structural.
• CONTRIBUTING.md and the README structural-check list now distinguish doc-completeness checks from behavioral enforcement — the runtime gate remains loop doctor + the contract's own verify scripts. The "gates are the bar" ground rule is unchanged: self_eval is still a hard CI gate; only the character of these three checks is clarified.

Note on the issue's stated proof gate: scripts/test_docs_claims.py only asserts receipts-claims-vs-artifacts, so it is vacuous for this change — the binding gates are self_eval 13/13 + test_self_eval + the full suite, all green (verified independently twice: 393 passed / 14 skipped; skips are the engine-extra env guards that run in CI).

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.