Plan vs. reality. Only 2 of 4 issues needed work — the other 2 were based on a misread of the matrix.
| Issue | Planned effort | Actual effort | Outcome |
|---|---|---|---|
| #1441 CC 2.1.117 adoption | 60 min | 40 min | 14 matrix entries, MIN bump, doc table, mcpServers audit |
| #1440 CHANGELOG ghost | 5 min | 2 min | Ghost removed |
| #1444 hook backfill (4 events) | 90 min | 0 min | Invalid premise — all 4 already registered |
| #1445 deprecated-ref sweep | 20 min | 0 min | Invalid premise — no actual drift |
agent_mcp_servers_main_thread — parallels 2.1.116's agent_hooks_main_threadfork_subagent_env_var (CLAUDE_CODE_FORK_SUBAGENT=1)plugin_install_auto_deps / marketplace_install_auto_depsmanaged_settings_marketplace_enforcemodel_pin_persists_restart + resume_large_session_summaryotel_command_attrs + otel_effort_attropus_46_sonnet_46_default_high — real cost change for Pro/Maxadvisor_tool_experimental + retention_expands_tasks_backupsnative_bfs_ugrep + opus_47_context_window_fix#1444, #1449, #1453 were all filed based on the brainstorm's misread of cc-version-matrix.ts: having a matrix entry means the feature is TRACKED in compatibility, not that there's no hook. All 27 hook events actually have at least one registered handler.
Tier A was scoped at ~2-3 hours; real work took ~45 min because 2 of 4 issues evaporated on verification. This is exactly the kind of error the brainstorm should have caught by grep'ing hooks.json before filing. Lesson for /ork:brainstorm: verify claims against current code state, not against mental model.
cd src/hooks && npx vitest run 7766/7766 pass npm run build ✓ npm run test:skills + test:agents ✓ ./bin/validate-counts.sh PASSED
Tier B (#1446 MCP audit, #1447 Monitor wiring, #1448 OTEL analytics) + Tier C (#1450 PushNotification, #1451 release cadence, #1452 property-based testing) still open as originally scoped. Test-infra follow-ups #1436 and #1438 also still open.