.forgejo/workflows/service-image.yaml builds Containerfile from the pushed
commit's tarball and pushes forgejo.coulomb.social/coulomb/informed-decision
:main-<short sha>, then reports the immutable digest. Before pushing it
checks the zlib floor, the runtime smoke (tools/zlib_overlay/check_runtime.py)
and that the installed modules equal the commit's source.
Containerfile builds the wheel in-image (hash-pinned setuptools in
requirements-build.lock) and installs the zlib overlay's hash-pinned,
Alpine-signed zlib-1.3.2-r1 package offline, failing below 1.3.2-r1, so the
live image's zlib fix cannot regress. Overlay workflows are unchanged.
Local build: 13/13 container checks, zlib 1.3.2-r1, source manifest match.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 5502@bnt-lap001
Assistant-Session: 68b3b02f-5d0f-4e71-8421-796f62e50196
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
