feat(orchestrator,tests): add CLI integration tests and refine orchestrator

Add tests/cli_integration.rs (+88 lines) with end-to-end CLI workflow
validation. Refine orchestrator pipeline coordination. Rotate 6 stale
beads history snapshots and sync tracker.

Co-Authored-By: Claude <noreply@anthropic.com>

Author: Dicklesworthstone

…y/issues.20260312_152532_783112752.jsonl …y/issues.20260315_153547_008845876.jsonl.beads/.br_history/issues.20260312_152532_783112752.jsonl renamed to .beads/.br_history/issues.20260315_153547_008845876.jsonl .beads/.br_history/issues.20260312_152532_783112752.jsonl renamed to .beads/.br_history/issues.20260315_153547_008845876.jsonl

@@ -202,6 +202,7 @@
 {"id":"bd-rjc","title":"U2 core follow-up: root-cause concurrent write-path sum drift/corruption","description":"While executing bd-2im with rch-offloaded runs, ci_scale remains unstable despite runtime-containment and retry hardening in harness. Evidence on 2026-02-25: (1) default mix run failed with final_sum 1024208 vs expected_sum 1024211; (2) transfer-only diagnostic failed with sum_delta=0 but final_sum 1024016 vs expected 1024000; (3) deposit-only diagnostic failed with final_sum 1024911 vs expected 1024943; (4) optional rusqlite verifier read hit malformed schema/rootpage on one run, indicating potential on-disk structural inconsistency under concurrent writer stress. Scope: isolate and fix root cause in fsqlite-core/pager/mvcc commit path (lost update/partial apply/corruption), add deterministic reproducer, then restore strict ci_scale invariant closure.","notes":"Support slice (PeachMaple, 2026-02-26): independent rch reproduction in /data/projects/frankensqlite using cargo test package fsqlite-e2e, test file bd_3plop_5_ssi_serialization_correctness, case ssi_serialization_correctness_ci_scale, with nocapture enabled, passed on this run (1 passed, 0 failed, about 0.79s). Symptom appears intermittent/non-deterministic; next cut should run repeated stress iterations with fixed seeds and contention profiling to estimate failure probability and isolate trigger conditions.","status":"closed","priority":0,"issue_type":"bug","assignee":"PinkGoose","created_at":"2026-02-25T22:10:57.920906200Z","created_by":"ubuntu","updated_at":"2026-02-27T22:04:55.982709839Z","closed_at":"2026-02-27T22:04:55.982640319Z","source_repo":".","compaction_level":0,"original_size":0,"labels":["cross-repo","frankensqlite","mvcc","quality"],"comments":[{"id":19,"issue_id":"bd-rjc","author":"Dicklesworthstone","text":"Support slice (EmeraldSparrow): added deterministic hot-row sum/integrity repro in /data/projects/frankensqlite at crates/fsqlite-e2e/tests/bd_rjc_deterministic_hot_row_sum_probe.rs. Validation: rch exec -- cargo test -p fsqlite-e2e --test bd_rjc_deterministic_hot_row_sum_probe -- --nocapture (pass: 1 passed, 0 failed).","created_at":"2026-02-26T02:58:44Z"},{"id":20,"issue_id":"bd-rjc","author":"Dicklesworthstone","text":"Support slice (EmeraldSparrow): hardened MVCC finalize path in /data/projects/frankensqlite/crates/fsqlite-mvcc/src/begin_concurrent.rs to preserve commit-index update + lock release when prepared session is missing/inactive during finalize. Added regression: test_finalize_releases_locks_and_updates_commit_index_when_session_missing. rch validation passed for new mvcc unit test and bd_rjc_deterministic_hot_row_sum_probe e2e test.","created_at":"2026-02-26T03:09:59Z"},{"id":21,"issue_id":"bd-rjc","author":"Dicklesworthstone","text":"Validation update (EmeraldSparrow, offloaded with rch): cargo check -p fsqlite-mvcc -p fsqlite-e2e passed. Clippy on touched crates surfaced unrelated pre-existing warnings/errors outside my edits: crates/fsqlite-vfs/src/unix.rs (useless_conversion) and crates/fsqlite-parser/src/semantic.rs (collapsible_match).","created_at":"2026-02-26T03:19:42Z"},{"id":26,"issue_id":"bd-rjc","author":"IcyMountain","text":"Support slice complete in /data/projects/frankensqlite/crates/fsqlite-e2e/tests/bd_rjc_deterministic_hot_row_sum_probe.rs: added tx_witness table writes in the same transaction as each hot-row increment (PRIMARY KEY(worker_id,round)), plus witness count/sum assertions against final value. Offloaded repro now fails deterministically with strong evidence of schema-layer corruption: expected at least one sqlite_master row but got none, while final_value=480, committed=480, witness_count=480, witness_delta_sum=480, retries=65. This indicates logical increments are conserved yet sqlite_master visibility/rootpage state is corrupted after concurrent commit stress. rch check: cargo check -p fsqlite-e2e PASS; rch test command consistently FAILS at sqlite_master assertion.","created_at":"2026-02-26T04:56:16Z"},{"id":29,"issue_id":"bd-rjc","author":"MagentaIvy","text":"Starting non-overlap support slice: repeated rch reproductions in /data/projects/frankensqlite (ci_scale + deterministic hot-row probe) to quantify intermittency and capture consistent failure signatures. No code edits planned in this slice unless requested.","created_at":"2026-02-26T04:57:36Z"},{"id":30,"issue_id":"bd-rjc","author":"MagentaIvy","text":"Repro batch complete (rch, /data/projects/frankensqlite): ci_scale test passed 3/3 while deterministic hot_row_probe failed 3/3. Failure signature remained stable: sqlite_master unexpectedly empty while value conservation still holds (final_value=480, committed=480, witness_count=480, witness_delta_sum=480). Observed retries in failing runs: 158, 101, 71. Batch log: /tmp/bd-rjc-repro-20260225_235810.log.","created_at":"2026-02-26T05:12:09Z"},{"id":39,"issue_id":"bd-rjc","author":"CloudyCave","text":"Observed while validating bd-194 from franken_whisper via rch: compile currently fails in /data/projects/frankensqlite/crates/fsqlite-vdbe/src/engine.rs with E0599 (CursorBackend::index_seek_prefix missing) and E0425 (record_prefix_all_non_null not found). This blocks franken_whisper cargo test execution before test bodies run.","created_at":"2026-02-27T03:56:11Z"},{"id":40,"issue_id":"bd-rjc","author":"Dicklesworthstone","text":"ROOT CAUSE CONFIRMED + FIX VALIDATED (AzurePond). Race in resolve_autocommit_txn(): plan_concurrent_commit() called OUTSIDE commit_write_mutex, while advance_commit_clock() inside it. Another connection can advance next_commit_seq between plan and finalize, corrupting commit_index. FIX: moved plan_concurrent_commit() inside commit_write_mutex in connection.rs:3696-3757. Also removed stale 'mut ok' after restructure. RESULTS: (1) bd_rjc_deterministic_hot_row_sum_probe: PASS, (2) concurrent_persist_five_threads: PASS (was FAILED), (3) concurrent_persist_10_threads: PASS (was FAILED), (4) clippy -D warnings: PASS. Remaining 2 ensure_column_exists failures are a separate bug (OpenRead on table rebuild root page).","created_at":"2026-02-27T22:04:29Z"}]}
 {"id":"bd-rxn","title":"Add edge-case tests for InsanelyFastPilot and DiarizationPilot","description":"InsanelyFastPilot and DiarizationPilot have basic happy-path tests but lack edge-case coverage matching WhisperCppPilot (bd-vup). Add: zero-duration, very short duration, extremely long duration, empty/edge config, boundary conditions. Files: src/backend/mod.rs.","status":"closed","priority":3,"issue_type":"task","assignee":"EmeraldSparrow","created_at":"2026-02-26T02:34:03.068719796Z","created_by":"ubuntu","updated_at":"2026-02-26T03:23:29.017490813Z","closed_at":"2026-02-26T03:23:29.017465796Z","close_reason":"edge-case tests present and validated via rch-offloaded pilot test groups","source_repo":".","compaction_level":0,"original_size":0,"labels":["enhancement"],"comments":[{"id":22,"issue_id":"bd-rxn","author":"Dicklesworthstone","text":"Validated existing bd-rxn edge-case coverage in src/backend/mod.rs for InsanelyFastPilot and DiarizationPilot (zero, single-ms, extremely long, boundary/contiguity/cycle/config edges). Offloaded verification: rch exec -- cargo test --lib insanely_fast_pilot_ -- --nocapture (12 passed) and rch exec -- cargo test --lib diarization_pilot_ -- --nocapture (17 passed).","created_at":"2026-02-26T03:23:28Z"}]}
 {"id":"bd-sfvm","title":"Refresh FRANKENTUI methodology robot event list","description":"docs/FRANKENTUI_METHODOLOGY.md still says robot mode event types are only run_start, stage, run_complete, run_error. Update it to describe the current robot contract including backends.discovery, health.report, routing_decision via robot routing-history, and speculation events.","status":"closed","priority":2,"issue_type":"task","assignee":"CyanAnchor","created_at":"2026-03-12T11:02:29.929743485Z","created_by":"ubuntu","updated_at":"2026-03-12T11:02:55.884173791Z","closed_at":"2026-03-12T11:02:55.880130041Z","close_reason":"Updated docs/FRANKENTUI_METHODOLOGY.md so its robot mode event list reflects the current contract: lifecycle, diagnostics, routing-history, and speculation events.","source_repo":".","compaction_level":0,"original_size":0}
+{"id":"bd-ssv2","title":"Honor optional pipeline stages with *.skip events","description":"Docs say VAD/diarize and other optional stages should be skipped when not requested, but orchestrator currently executes Align/Punctuate/Diarize unconditionally and does not emit skip events. Scope: gate optional stages off request/config, preserve deterministic eventing, add tests for non-diarize/non-vad runs.","status":"blocked","priority":1,"issue_type":"task","created_at":"2026-03-15T15:35:46.972068405Z","created_by":"ubuntu","updated_at":"2026-03-15T15:43:05.745168573Z","source_repo":".","compaction_level":0,"original_size":0}
 {"id":"bd-vgh5","title":"Include FW-JSON in parity error-code inventory","description":"docs/FEATURE_PARITY.md lists per-variant FW-* error codes but omits FW-JSON even though src/error.rs exposes it and README documents it. Update the parity docs so the inventory is complete and matches the implementation.","status":"closed","priority":3,"issue_type":"task","assignee":"CyanAnchor","created_at":"2026-03-12T10:59:00.649184067Z","created_by":"ubuntu","updated_at":"2026-03-12T10:59:21.883036521Z","closed_at":"2026-03-12T10:59:21.879759150Z","close_reason":"Added FW-JSON to the per-variant error-code inventory in docs/FEATURE_PARITY.md so the parity docs match src/error.rs and README.","source_repo":".","compaction_level":0,"original_size":0}
 {"id":"bd-vup","title":"Add WhisperCppPilot error scenario tests","description":"WhisperCppPilot has 10 happy-path tests but no error injection tests. Add tests for: missing model path handling, zero-duration audio, extremely long audio, timeout scenarios. File: src/backend/mod.rs, WhisperCppPilot section.","status":"closed","priority":2,"issue_type":"task","assignee":"GentleMink","created_at":"2026-02-26T02:05:15.715799046Z","created_by":"ubuntu","updated_at":"2026-02-26T02:28:08.198357111Z","closed_at":"2026-02-26T02:28:08.198338015Z","close_reason":"Added 9 error/edge-case tests for WhisperCppPilot: zero-duration (empty output), 1ms audio (single segment), 24hr extremely long duration (17280 segments), empty model path, zero threads, translate flag storage, phrase cycle wrap, contiguous segment boundaries, exact 5000ms boundary. All 2804 tests pass, clippy clean.","source_repo":".","compaction_level":0,"original_size":0,"labels":["enhancement"]}
 {"id":"bd-wlnf","title":"Harden rch quality-gate retries against sync pollution and local fallback","status":"closed","priority":1,"issue_type":"task","created_at":"2026-03-12T15:23:52.916185437Z","created_by":"ubuntu","updated_at":"2026-03-12T15:24:53.472477890Z","closed_at":"2026-03-12T15:24:53.465209731Z","close_reason":"Completed","source_repo":".","compaction_level":0,"original_size":0}