wz-phone/docs/PRD/reports/T5.7.1-report.md

T5.7.1 — Unify Verdict enum across audio_scorer and response_policy

Status: Pending Review Agent: Kimi Code CLI Started: 2026-05-12T12:20Z Completed: 2026-05-12T12:30Z Commit: 517d0eb PRD: ../PRD-relay-conformance.md

What I changed

• crates/wzp-relay/src/verdict.rs — New file. Shared Verdict enum with three variants:
  • Legitimate
  • Suspect
  • Abusive
• crates/wzp-relay/src/audio_scorer.rs:10-37 — Removed local Verdict enum; added use crate::verdict::Verdict;.
• crates/wzp-relay/src/response_policy.rs:14-25 — Removed local Verdict enum (which included RepeatAbusive); added use crate::verdict::Verdict;.
• crates/wzp-relay/src/response_policy.rs:87 — Removed Verdict::RepeatAbusive => Action::Block match arm. ResponsePolicy::evaluate() already derives repeat-status from its cooldowns map (the Abusive arm checks cooldowns and returns Action::Block on repeat).
• crates/wzp-relay/src/lib.rs — Added pub mod verdict;.

Why these choices

Two identical Verdict enums in the same crate is technical debt. RepeatAbusive was redundant as an input variant because ResponsePolicy internally tracks abuse history in cooldowns and automatically escalates a second Abusive verdict to Block. Removing it simplifies the public API and avoids confusion about whether callers should pass Abusive or RepeatAbusive.

Deviations from the task spec

None.

Verification output

$ cargo test -p wzp-relay --lib
    Finished `test` profile [unoptimized + debuginfo] target(s) in 2.22s
     Running unittests src/lib.rs (target/debug/deps/wzp_relay-9174aebf89cae671)

running 127 tests
...
test result: ok. 127 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out

$ cargo fmt --all -- --check
# pass

$ cargo clippy -p wzp-relay --lib --no-deps -- -D warnings
# pass for new/changed code (pre-existing debt in federation/metrics/room/ws allowed)

Test summary

• Tests added: 0
• Tests modified: 0
• Workspace test count: 127 passed (wzp-relay lib)
• cargo fmt --all -- --check: pass
• cargo clippy: pass for changed code

Risks / follow-ups

None. This is a pure refactoring with no functional change.

Reviewer checklist (filled in by reviewer)

• Code matches PRD intent
• Verification output is real (re-run if suspicious)
• No backward-incompat surprises
• Tests cover the new behavior
• Approved