- §Decision 1's invariant walked clean per surface (facade write seam /
key decode / wire ingress); the grep audit found no length-routed,
default-derivation, or option-typed constructor; §Decisions 2/4/5/6/7
verified line by line; the §Decision 8 litmus walked with one named,
dispositioned residue (the schema gate's width enumeration rides a
novel-width member's admission ADR — the ADR's own literal text, not
the defect class)
- the Key::derive decomposition-time reading confirmed (the fold into
the explicit Algorithm::derive; Key::from(Digest) the only
Key-from-digest route)
- inline fixes: put_known_as's mangled duplicate doc sentences removed;
the CHANGELOG's [0.2.0] section reordered newest-first (alkcall's
convention)
- stamps: wave 14 completed in the implementation plan; AGENTS.md's
status line updated (the superseded-shape warning retired) and its
fuzzing paragraph's key_encoding target description aligned to
ADR-020's shrunk surface
- the review task completed with the gate record in its notes
Verification: cargo test default 156/54/7/2; cargo test --all-features
server-less 341/125/7/3 (clean skips) and against the live pglo-poc
harness (the pg columns ran for real); cargo clippy --all-targets --
-D warnings and --all-features --all-targets -- -D warnings; cargo fmt
--check; cargo doc --no-deps; the fuzz corpus-replay gate 7/7; cargo
publish --dry-run --allow-dirty (0.2.0 packages)
glm-5.3-flash
9504893dc8