docs(aeoess-aps): revocation mapping, one TRACE mechanism at two levels of evidence, nothing emitted (#140)

[aeoess]

Revocation mapping for the aeoess-aps integration, following the questions in #140: one TRACE mechanism at two levels of evidence, with the mapping stated per surface and nothing emitted on the APS side that TRACE does not define. README gains a Revocation section; integration.yaml gains the tested_against note. The pointer to trace-spec#246 is filled.

[github-actions]

🔴 Contributor Check: HIGH

Check	Result
Profile	HIGH
Credential	LOW
Overall	HIGH

Automated check by AgenTrust Contributor Check.

[imran-siddique]

This lands the ruling and adds something I did not ask for that is better than what I did ask for.

The reframing is right: one mechanism at two levels of evidence, entry-ID-scoped where there is an inclusion proof and binary on the key where there is not, rather than a released surface and a prospective one. Keeping the per-surface columns while changing the frame above them was the correct call.

The tested_against comment is the part worth pointing at:

trace-spec#236 (merged 2026-08-29, unreleased) changes the RevocationStore path: a callable returning a non-bool now fails closed instead of reading as not revoked. The revocation observations in README were made against 0.9.0 and are re-run on the release that carries #236.

Three things right about that, in order of how often they get skipped. It records that a result has a known expiry rather than presenting it as current. It sits next to tested_against: "0.9.0", so a reader who checks the pin finds the caveat rather than having to know to look for it. And it commits to the re-run without pretending it has happened.

The alternative most contributors take is to quietly leave observations that were true when made and are now stale, which is exactly the failure trace-spec#241 is open about at the package level.

All fourteen substantive checks green, including validate, which is the one that catches a drifted generated index when integration.yaml changes, plus links, ruff and four conformance matrices.

Merging. Closing #140: all three questions are answered, the mapping reflects the answers, and the divergence you found is filed as trace-spec#246 with your report credited.