spec(port): port the commit signature forgery guard (Refs #8883)

[dmitrii-f-t27]

The third CR-CHAT-03 ring target (W3, tracker #8883, epic gHashTag/trios#1812): port of crates/trios-chat/rings/CR-CHAT-03/src/commit_signature.rs (Wave-24 Lane A, L-CHAT-3-csig, R-CHAT-3, RFC 9420 6.1/12).

Every MLS Commit is signed by its committer leaf key. The forgery gate pins the binding invariants before the caller delegates to the real signature check. Rules, in the original's fixed order:

1. sig_blob non-empty AND non-zero (lazy-verifier trap),
2. group_id agreement (splice rejection),
3. from_epoch == current_epoch,
4. committer currently a member,
5. sig_pubkey == stored leaf-public (stale pre-rotation key rejected),
6. ops_hash == locally computed hash (cross-commit splice).

Carried as the D5 verdict — the one variant with payload (EpochMismatch) carries the current and claimed epochs; the rest are unit codes.

Adaptations (t27 conventions): transcript/view structs → flat parameters, the signature blob as a plain slice; the view's Option<[u8; 32]> committer key → member flag + stored key (sentinel convention); the transcript's sender_leaf rides along unread by every rule, exactly as in the original. The original's tenth test is a tautology green summary (a count asserted against itself) — replaced with a live canonical gate call, the WPS-10 precedent, so no test passes vacuously.

All 10 original CSF tests transcribed with exact asserts (CSF-09 is the original's own check-order invariant: with every other field also wrong, the empty signature is still the answer), plus helper tests and a full rule-order ladder stacking every later violation under each earlier rule (13 total).

Checks (conventions 0-14): t27b test 13 passed 0 failed, 0 blockers; t27c test-report 13/13, 0 vacuous; gen-rust output compiles under rustc (0 'static str, 0 Vec<const); coverage Untested 0; ASCII only; seal saved, all hashes MATCH; ledger row added to docs/reports/t27b_expectations.json.

Refs #8883

[dmitrii-f-t27]

All required checks green (validate, check-linked-issue, parse-ratchet, t27b-native-linux, t27b-native-macos); cli-tri / duplicate-bodies / t27b-native-ratchet are the known non-required reds. Steward READY. Third CR-CHAT-03 target (3/47 T-PORT) — requesting owner merge when convenient, Refs #8883.

[dmitrii-f-t27]

Superseded by #10175: the remaining CR-CHAT targets now land as one port-wave batch per specs/policy/pr_budget.t27 (#10067/#10074), which asks for one PR per ring batch instead of one per file. The spec body, seal and ledger row in this PR are unchanged there; the batch also carries one self-consistent duplicate-body bless. Closing this one-file PR keeps the merge queue and the Actions queue small.