glm-5.3-flash
d88e148fdc
alkcall 0.5.0 landed the review 006 remediation (filed from this crate's Phase 0 pass): - E-01/N-1: channel-open establishment phase (ADR-049) — register_openable_with_establisher, bounded await, teardown-on- failure, typed channel:open_failed reply (reason in details); ChannelClient::open_channel returns ChannelOpenError - E-02: OperationSpec.description disclosed via discovery - E-03/E-04/N-2: teardown-race log, early-arrival bound docs, count accessor alktunnels updates: - Cargo.toml: alkcall 0.4.0 -> 0.5.0 (lockfile resolves 0.5.0) - AGENTS.md convention 11: pin note updated; ADR-049 marked load-bearing (tunnel opens use register_openable_with_establisher) - phase-0-findings OQ-TN-09: RESOLVED by ADR-049 — the in-band control-frame hunch is superseded (establishment is the open op's reply, not a stream frame); residual for Phase 1 narrowed to half-open semantics; v1 needs no mid-stream control frame - phase-0-findings OQ-TN-08: per-op metadata residue resolved by E-02 (description field; OQ-40 enumerate half stays deferred) - tun2proxy prior-art + alktty-pattern sections: frame-vocabulary notes updated to point at the ADR-049 supersession - upstream posture paragraph updated: review 006 is the working proof of the ask-early rule Verification: cargo test (0 tests, ok), clippy --all-targets -D warnings, fmt --check, wasm32 check, doc --no-deps — all clean