The vocabulary mapping this path was waiting on is not coming: gate-house
rejected it in GH-DEC-2026-008, because a translation can be confidently
wrong and fails open by accepting a claim approved for a different action.
The stronger option arrived instead, and both halves are enforced here.
flex-auth published binding.approval_binding_digest (FLEX-DEC-2026-007) to
fix the circularity this repo reported: a pdp_digest recorded at issue time
can never equal the request_digest of the request that carries the claim in
its hashed context, so with GH-DEC-2026-008 requiring that equality, destroy
would have failed closed forever on a check no correct record could pass.
- authorization.approval_binding_digest implements the published exclusion
rule, including Go's context,omitempty behaviour when stripping empties
the context; digest_material drops an empty context for the same reason.
- validate_decision_envelope recomputes the field rather than trusting it,
refuses a claim-bearing request whose decision records none, and compares
the claim's digest from step 1 against it -- never against request_digest,
which still covers the claim so it stays a sound replay identity.
- validate_approval_claim requires binding.pdp_path true before using
pdp_digest at all. Path intent is never inferred from a digest that
happens to be present; pre-schema-v3 approvals carry pdp_path false
regardless of any digest they hold.
Replay fixtures re-vendored from dd3ce4c. The destroy pins moved a second
and final time; approval_binding_digest did not, which is the point. The
fixture now demonstrates the property instead of asserting it: we rederive
fa07becf... from its own request through our canonical implementation,
proving we hash the same material flex-auth does rather than pinning a
constant we cannot reproduce.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01E4tNMAYcSQmZWUE4wqP4ij
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 715726@bnt-lap001
Assistant-Session: 80a42b32-cba6-4b23-8be0-68819b1a6092
Implements step 2 (access-engine POST /v1/check) and wires the whole
GH-DEC-2026-003 sequence together, then proves it against a live throwaway
OpenBao rather than only unit-level fakes.
- decision_check.check_decision performs the PDP call; an unreachable or
non-200 PDP raises, since silence is never permission.
- approval_consume.authorize_action coordinates steps 1 and 2 and returns an
AuthorizedAction. Both steps build the same CheckRequest via a shared
_expected_request, since two descriptions of the action cannot produce
corresponding digests.
- apply_unreachable_engine_stance takes authorized=. The published map
defines fail_closed as no side effect WITHOUT a durable decision record,
so holding a validated one means the residue does not apply. Not a bypass:
both steps must have succeeded and CAS consume still precedes OpenBao.
Unconfigured still returns None and fails closed.
The end-to-end test caught one more instance of the cross-vocabulary bug: a
leftover comparison of the claim's binding.action against ours. The claim
says secrets.kv.destroy where we say destroy, so it would have failed against
every real claim. Removed; the tie is pdp_digest.
Integration coverage asserts PIP-then-PDP ordering, that consume is the last
step before the backend, and that an unreachable PDP, denied decision,
invalid claim, missing pdp_digest, consume conflict and action mismatch each
stop before OpenBao. 284 tests pass; production still fails closed.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01M65ovP3eiiPHubibvWs9mD
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 393550@bnt-lap001
Assistant-Session: 4bb359f9-1f12-4410-9e76-079cf23c82e4
