approval-engine met the condition. Verified here rather than taken on report:
their docs/approval-claim.md carries "Presentation exclusion — GH-DEC-2026-015
§4" in normative language, and I ran
tests/test_claim_contract.py::test_presentation_changes_cannot_change_the_approved_act
myself — 1 passed. That test pins the digest input set from BOTH sides, and the
narrowing half is what makes it real: without it a digest over four fields, or
over a constant, would pass the widening half perfectly.
view_hash now carries binding.digest and the act-scope is no longer
independently canonicalized here, so the act has exactly one canonicalization
computed by the layer that owns it. approval_binding_digest is validated for
shape and refused without its approval id — it is carried, never computed.
The three published vectors are unchanged: they do not carry the new key, so
pick omits it. Asserted, not assumed.
The cycle condition did not disappear, its protection moved — from refusing
nesting to approval-engine's normative exclusion. layer.yaml carries it as
cycle_condition with a test, so a future widening meets a rule rather than
silence.
One thing not assumed. Both gate-house and approval-engine said our binding
slice canonicalizes principal and target, two of their five fields. target
plainly is act material and is now dropped. But their principal is the party ON
WHOSE BEHALF the approval was issued, while ours is the person being BOUND — the
approver. Different roles, and dropping ours would remove who was shown this
from view_hash and gut the promise. Kept it, declared principal_role_overlap
open in layer.yaml, tested that changing the approver still moves view_hash, and
raised it rather than silently resolving it either way.
L0/L2 are unaffected: with no approval there is no digest to defer to, and
test_act_scope_still_binds_when_there_is_no_carried_digest pins that.
100 tests pass.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01V3W1dQG7GFFM9d94jFx7iR
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 1565372@bnt-lap001
Assistant-Session: 16bb2f25-b34c-49ef-8e94-5fec3567a568
Verified the three published hashes reproduce byte for byte before promoting
anything, then moved the schema, canonicalizer and vectors into governed assets.
history/20260909-initial-exploration/ is untouched and stays the provenance
record.
- schemas/, informed_decision/, tests/vectors/ populated; the reference
canonicalizer's ad-hoc __main__ block replaced by a real
`python -m informed_decision` entry point.
- tests/test_canonicalize.py — 20 tests, all green. Published vectors, all four
isolation properties, canonical-form round-trip, key sorting, and a provenance
test asserting the governed fixtures have not drifted from history/.
- docs/specs/EvidenceModel.md — the two hashes, the split and why it exists, the
four isolation properties, the presentation record, the bundle, and the
relationship to audit-core.
- pyproject.toml, Makefile.
One test of mine was wrong on first run: it scanned for ", " to assert no
insignificant whitespace, which fires on prose inside a brief. Replaced with a
canonical round-trip comparison, which is the property actually meant. The
canonicalizer was correct.
EvidenceModel leads with what the model does NOT claim — no proof of
comprehension, no proof of reading (deliberately, since the alternative is
surveillance), no survival of a compromised surface, and audit-core's inherited
bound that a hash chain cannot prove a record was never sent.
T06 stays progress: the SCOPE.md rewrite is gated on the T02 ruling.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01V3W1dQG7GFFM9d94jFx7iR
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 1565372@bnt-lap001
Assistant-Session: 16bb2f25-b34c-49ef-8e94-5fec3567a568
