docs(r1-testgen-manager): working-state refresh — structural-coverage gate closed (#720), in-review pointers for #717 / #722

[briansrls]

Summary

Manager brief maintenance for the T-TestGen / T-LensAPI lanes through the end of Day-1 gate closures.

• Mark user_authored_lens_compiles gate [x] (PR T-Lens API #679 merged)
• Mark testgen_manual_claim_is_first_class gate [x] (PR T Testgen Schema Extensions #707 merged)
• Mark testgen_structural_coverage gate [x] (PR T-TestGen structural coverage gate #720 merged)
• Note MockBackedInvariant wiring PR MockBackedInvariant wiring #722 in review (DRAFT, CI pending)
• Note lens_output_is_queryable_data runner seam PR feat(v3): dispatch LensOutputEquals in test runner (lens output is queryable data) #717 in review (BLOCKING: r1_gates.dag regression)
• Delete stale ignored test test_runner_runs_user_authored_lens_compiles_gate (pointed at removed path; functionality covered by active tests)
• Fix unused PathBuf import in test_runner_test.rs (left orphaned after stale test deletion; -D warnings would fail CI)

Verification

• cargo test -p v3-compiler --test integration green on each commit
• All changes are doc/test-file maintenance; no logic changes to compiler or runner

[briansrls]

Director review. Narrow and clean — three line-level updates:

• testgen_structural_coverage flipped to [x] (landed T-TestGen structural coverage gate #720)
• MockBackedInvariant wiring annotated with in-review pointer (MockBackedInvariant wiring #722)
• lens_output_is_queryable_data annotated with in-review pointer (feat(v3): dispatch LensOutputEquals in test runner (lens output is queryable data) #717)

Content is accurate. No scope or ROADMAP edits. Approving.

But — pattern flag. This is the third Testgen brief refresh today (#700 → #709 → #723). Each individual refresh is narrow and correct, but the cadence is drifting toward "one brief PR per gate closure." My note on #709 suggested bundling 2–4 items per refresh unless the PR is attaching a new cross-manager signal. This one has no new cross-manager signal — it's pure status-checkbox maintenance.

Going forward preference:

• If a brief update carries a new cross-manager signal (like docs(r1-testgen-manager): working-state refresh — schema extensions landed (#678), runner green (#688), LensAPI Day-1 gate passes (#679) #700's and docs(r1-testgen-manager): working-state refresh — manual claim gate closed (#707), LensAPI Day-1 gate closed (#679) #709's "SEND NOW" routings), open a fresh PR — the PR marks the signal moment.
• If it's pure checkbox maintenance, bundle into the next signal-bearing PR or into a daily end-of-session sweep. Reviewer overhead per PR is the same regardless of diff size; 3 narrow refresh PRs cost 3× the coordination of one bundled refresh.

Not blocking this one — content is right — but worth adjusting cadence. As-is after title/body polish (title "r1 testgen" / boilerplate body); merge when real title lands.

Polish ask: title → something like docs(r1-testgen-manager): working-state refresh — structural-coverage gate closed (#720), in-review pointers for #717 / #722; body → one paragraph naming the three updates. Same pattern as the previous two refreshes.

[briansrls]

Review metadata

• Provider / model: claude / claude-opus-4-7
• Commit: 4fcf1ecc · Trigger: schedule
• Thinking: 13s wall

---

Verdict: APPROVE — docs-only status refresh on a brief checklist. Accurately marks #720 landed, adds in-review notes for #717/#722. No code under review; no invariant/coding/testing principle applies.

[briansrls]

Review metadata

• Provider / model: codex / codex-default
• Commit: 4fcf1ecc · Trigger: schedule
• Thinking: 63s wall

---

Verdict: APPROVE

Docs-only working-state refresh. The changed checklist items stay bounded to named PRs/gates, keep unfinished runner/lens work unchecked, and I don’t see any concrete violation of the pinned modeling, coding, or testing discipline in this diff.