Thoughtbox

{"id":"thoughtbox-1t3","title":"Fix Phase 2 eval system post-review","status":"closed","priority":1,"issue_type":"bug","created_at":"2026-02-12T16:06:03.570331-06:00","updated_at":"2026-02-12T16:08:52.627464-06:00","closed_at":"2026-02-12T16:08:52.627464-06:00","close_reason":"All review issues resolved: compilation blocker, detach safety, deduplication, error handling","created_by":"glassBead"} {"id":"thoughtbox-25q","title":"Add Channels and Consensus tabs to observatory","status":"open","priority":2,"issue_type":"feature","created_at":"2026-02-09T12:52:43.5941-06:00","updated_at":"2026-02-09T12:52:43.5941-06:00","created_by":"glassBead"} {"id":"thoughtbox-2ln","title":"Master behavioral test suite for all 60+ Thoughtbox operations","status":"closed","priority":1,"issue_type":"feature","created_at":"2026-02-07T01:33:22.117949-06:00","updated_at":"2026-02-07T01:59:07.843157-06:00","closed_at":"2026-02-07T01:59:07.843157-06:00","close_reason":"All 116 behavioral tests written across 10 files, master index created, live-tested against Docker server, arg shapes verified and corrected","created_by":"glassBead"} {"id":"thoughtbox-2te","title":"Rename MANAGER → COORDINATOR + add manager capability flag + fix server bugs","status":"closed","priority":1,"issue_type":"feature","created_at":"2026-02-07T17:17:55.84278-06:00","updated_at":"2026-02-07T17:31:15.604148-06:00","closed_at":"2026-02-07T17:31:15.604148-06:00","close_reason":"All 4 parts complete: MANAGER→COORDINATOR rename, manager flag, ESM fix, once-per-session priming fix","created_by":"glassBead"} {"id":"thoughtbox-2wx","title":"Observatory: Connect to persistent storage for historical sessions","status":"closed","priority":1,"issue_type":"feature","created_at":"2026-02-07T14:28:11.207202-06:00","updated_at":"2026-02-07T14:33:51.564596-06:00","closed_at":"2026-02-07T14:33:51.564596-06:00","close_reason":"Implemented: adapter, server plumbing, UI, tests all passing","created_by":"glassBead"} {"id":"thoughtbox-308","title":"Profile priming injects on every thought call — should be once per session","description":"gateway-handler.ts lines 504-516: getProfilePriming() is called inside handleThought(), meaning every single thought call appends the full mental model contents. Should only prime once per session (e.g., on first thought or on cipher load). This bloats every thought response with redundant content.","status":"closed","priority":1,"issue_type":"bug","created_at":"2026-02-07T08:53:17.324974-06:00","updated_at":"2026-02-07T17:31:18.896095-06:00","closed_at":"2026-02-07T17:31:18.896095-06:00","close_reason":"Fixed in commit 7828402 - sessionsPrimed set guards once-per-session priming","created_by":"glassBead"} {"id":"thoughtbox-32q","title":"Wire clearSession() into transport close lifecycle","status":"open","priority":3,"issue_type":"bug","created_at":"2026-02-09T12:37:09.761325-06:00","updated_at":"2026-02-09T12:37:09.761325-06:00","created_by":"glassBead"} {"id":"thoughtbox-34w","title":"Export system broken: require is not defined (ESM/CJS mismatch)","description":"The export system (both auto-export on session close and manual session.export with format=json) fails with 'require is not defined'. This is a CommonJS require() call in an ESM module context. Affects 9 behavioral tests: TB-008 through TB-015 (linked node structure verification) and SS-009 (JSON export). Auto-export warning appears on every session close: 'Auto-export failed: require is not defined. Session remains open to prevent data loss.' The export codepath in the gateway uses require() but the server runs as ESM.","status":"in_progress","priority":1,"issue_type":"bug","created_at":"2026-02-08T05:50:44.426661-06:00","updated_at":"2026-02-08T06:05:54.359842-06:00","created_by":"glassBead"} {"id":"thoughtbox-3iw","title":"Module 3: agent-identity","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-05T19:38:23.773814-06:00","updated_at":"2026-02-05T19:42:25.431461-06:00","closed_at":"2026-02-05T19:42:25.431461-06:00","close_reason":"Closed","created_by":"glassBead"} {"id":"thoughtbox-3s6","title":"Implement /multi-agent-tdd slash command (8 modules, 67 tests)","status":"closed","priority":1,"issue_type":"feature","created_at":"2026-02-05T23:31:20.184431-06:00","updated_at":"2026-02-06T00:16:40.27495-06:00","closed_at":"2026-02-06T00:16:40.27495-06:00","close_reason":"All 8 modules implemented: 67 tests passing across 8 test files. Hub regression (117 tests) clean. Build clean.","created_by":"glassBead"} {"id":"thoughtbox-42r","title":"Eval: thought content sent verbatim to LangSmith — review data sensitivity","description":"onThoughtAdded sends data.thought.thought as run output. If thoughts contain sensitive content, this is a data leak to LangSmith. Need to assess whether sanitization or opt-in is needed. Review finding P1.","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-12T16:17:07.204411-06:00","updated_at":"2026-02-12T21:10:48.567289-06:00","closed_at":"2026-02-12T21:10:48.567289-06:00","close_reason":"Fixed in Phase 2 review cleanup","created_by":"glassBead"} {"id":"thoughtbox-4fm","title":"HDD: Backfill agentops test coverage (20% → target 60%+)","status":"open","priority":2,"issue_type":"task","created_at":"2026-02-05T21:26:32.750006-06:00","updated_at":"2026-02-05T21:26:32.750006-06:00","created_by":"glassBead"} {"id":"thoughtbox-4j7","title":"M11: Wire hub-handler and server-factory","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-06T03:15:48.308252-06:00","updated_at":"2026-02-06T03:21:53.141549-06:00","closed_at":"2026-02-06T03:21:53.141549-06:00","close_reason":"Closed","created_by":"glassBead","dependencies":[{"issue_id":"thoughtbox-4j7","depends_on_id":"thoughtbox-ih7","type":"blocks","created_at":"2026-02-06T03:15:54.109312-06:00","created_by":"glassBead"}]} {"id":"thoughtbox-4sw","title":"Module 4: hub-tool-wiring","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-05T19:38:23.834269-06:00","updated_at":"2026-02-05T19:48:09.582891-06:00","closed_at":"2026-02-05T19:48:09.582891-06:00","close_reason":"Closed","created_by":"glassBead"} {"id":"thoughtbox-5gc","title":"Update hub tool description and schema in server-factory","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-07T04:05:01.255771-06:00","updated_at":"2026-02-07T04:09:24.441835-06:00","closed_at":"2026-02-07T04:09:24.441835-06:00","close_reason":"Hub tool description, schema, and callback wiring complete","created_by":"glassBead","dependencies":[{"issue_id":"thoughtbox-5gc","depends_on_id":"thoughtbox-cf6","type":"blocks","created_at":"2026-02-07T04:05:07.522098-06:00","created_by":"glassBead"}]} {"id":"thoughtbox-6ac","title":"M10: Harden gateway-handler.ts with per-session identity","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-06T02:27:48.715157-06:00","updated_at":"2026-02-06T02:33:25.305075-06:00","closed_at":"2026-02-06T02:33:25.305075-06:00","close_reason":"Closed","created_by":"glassBead"} {"id":"thoughtbox-6d2","title":"Implement hub_wait long-polling tool","status":"closed","priority":1,"issue_type":"feature","created_at":"2026-02-07T06:27:55.58612-06:00","updated_at":"2026-02-07T06:31:17.597886-06:00","closed_at":"2026-02-07T06:31:17.597886-06:00","close_reason":"Implemented hub_wait with 15 tests, all passing. WaitManager + hub handler integration + server factory schema.","created_by":"glassBead"} {"id":"thoughtbox-7rv","title":"Module 5: channel-resources","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-05T19:38:23.89037-06:00","updated_at":"2026-02-05T19:48:53.618669-06:00","closed_at":"2026-02-05T19:48:53.618669-06:00","close_reason":"Closed","created_by":"glassBead"} {"id":"thoughtbox-8ut","title":"Eval: inconsistent branch metric fallback order in dgm-fitness evaluator","description":"dgm-fitness.ts falls back branchingFactor→0 but session-quality falls back differently. Inconsistent metric extraction order across evaluators. Review finding P1.","status":"closed","priority":2,"issue_type":"bug","created_at":"2026-02-12T16:16:57.876422-06:00","updated_at":"2026-02-12T21:10:48.546848-06:00","closed_at":"2026-02-12T21:10:48.546848-06:00","close_reason":"Fixed in Phase 2 review cleanup","created_by":"glassBead"} {"id":"thoughtbox-914","title":"M10: Write per-session identity isolation tests","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-06T02:27:51.101733-06:00","updated_at":"2026-02-06T02:33:25.318102-06:00","closed_at":"2026-02-06T02:33:25.318102-06:00","close_reason":"Closed","created_by":"glassBead"} {"id":"thoughtbox-9a2","title":"ADR-001: MCP Hub - Collaborative Reasoning","description":"Build an MCP hub into thoughtbox that enables multiple MCP client agents to collaborate on structured reasoning using Git-like version control patterns.","notes":"HDD Session Metadata\nEpic ID: thoughtbox-9a2\nPhase 1 (Research): thoughtbox-9a2.1\nPhase 2 (Staging ADR): thoughtbox-9a2.2\nPhase 3 (Implementation): thoughtbox-9a2.3\nPhase 4 (Validation): thoughtbox-9a2.4\nPhase 5 (Decision): thoughtbox-9a2.5\n\nCurrent Phase: Phase 1 (Research)\nStatus: Agent working on research","status":"closed","priority":2,"issue_type":"epic","created_at":"2026-02-05T15:21:56.011873-06:00","updated_at":"2026-02-05T21:01:48.680933-06:00","closed_at":"2026-02-05T21:01:48.680933-06:00","close_reason":"Epic complete. All 5 HDD phases done: Research → Staging ADR → TDD Implementation (74 tests) → MCP Wiring → Docker E2E Validation → Decision: ACCEPT.","created_by":"glassBead"} {"id":"thoughtbox-9a2.1","title":"Phase 1: Research and Hypothesis Formation","description":"Research existing thoughtbox architecture, MCP protocol, Git-like version control patterns for reasoning. Form testable hypotheses.","notes":"Phase 1 Research Complete.\n\nKey findings:\n- Three-component architecture: Proxy (communication) → Hub (coordination) → Thoughtbox (reasoning)\n- MCP Protocol Tasks (2025-11-25 spec, experimental) is the backbone for hub task lifecycle\n- No MCP clients support Tasks — proxy bridges this gap via capability negotiation\n- Existing thought/branch/session primitives are the reasoning substrate (unchanged)\n- Gateway stays focused on per-agent context management — hub has its own tool\n- Observatory is visualization only — collaboration logic belongs in hub\n- Two roles: coordinator (first among equals) + contributor\n- Channels hold per-task message state as dynamically updating resources\n- Presence registry tracks connected agents\n- 6 hypotheses documented (H1-H6), all testable\n\nDeliverables:\n- staging/docs/adr/001-mcp-hub-research.md\n- staging/docs/adr/001-mcp-hub-diagrams.md","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-05T15:23:27.784606-06:00","updated_at":"2026-02-05T16:15:06.799415-06:00","closed_at":"2026-02-05T16:15:06.799415-06:00","close_reason":"Research complete. Architecture settled: Proxy/Hub/Thoughtbox. All findings and diagrams documented.","created_by":"glassBead","dependencies":[{"issue_id":"thoughtbox-9a2.1","depends_on_id":"thoughtbox-9a2","type":"parent-child","created_at":"2026-02-05T15:23:27.7852-06:00","created_by":"glassBead"}]} {"id":"thoughtbox-9a2.2","title":"Phase 2: Staging ADR Creation","description":"Create staging ADR with context, decision, hypotheses for MCP Hub collaborative reasoning design.","notes":"Phase 2 Staging ADR drafted: staging/docs/adr/002-mcp-hub-staging-adr.md\n\nCovers:\n1. Full data model (9 types: AgentIdentity, Workspace, WorkspaceAgent, Problem, Comment, Proposal, Review, ConsensusMarker, Channel)\n2. Hub operations (20 operations across 6 categories: identity, workspaces, problems, proposals, consensus, channels)\n3. Channel mechanics — MCP Resources with subscription support, URI scheme: thoughtbox://hub/{workspace}/channels/{problem}\n4. Proxy specification — capability handshake, 4-cell capability matrix, in-process Tasks adapter\n5. Merge thought semantics — synthesis on main chain, branch preservation\n6. Progressive disclosure — 3 access stages\n7. Storage layout — JSON files alongside existing session storage\n8. Module structure — 9 source files + test suite\n9. MCP Tasks mapping — immediate vs task-eligible operations, state mapping\n10. Hypothesis validation plan\n11. Migration/compatibility guarantees\n12. Open questions\n\nAwaiting user review.","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-05T15:24:13.679228-06:00","updated_at":"2026-02-05T20:42:07.013932-06:00","closed_at":"2026-02-05T20:42:07.013932-06:00","close_reason":"Staging ADR created at staging/docs/adr/002-mcp-hub-staging-adr.md, reviewed and approved","created_by":"glassBead","dependencies":[{"issue_id":"thoughtbox-9a2.2","depends_on_id":"thoughtbox-9a2.1","type":"blocks","created_at":"2026-02-05T15:24:55.044271-06:00","created_by":"glassBead"},{"issue_id":"thoughtbox-9a2.2","depends_on_id":"thoughtbox-9a2","type":"parent-child","created_at":"2026-02-05T15:24:13.679838-06:00","created_by":"glassBead"}]} {"id":"thoughtbox-9a2.3","title":"Phase 3: Implementation","description":"Implement MCP Hub according to ADR, write tests.","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-05T15:24:13.730409-06:00","updated_at":"2026-02-05T20:42:12.592063-06:00","closed_at":"2026-02-05T20:42:12.592063-06:00","close_reason":"11 hub modules implemented via TDD (74 tests), wiring layer complete, registerToolTask upgrade done. 117 tests pass, TS compiles clean.","created_by":"glassBead","dependencies":[{"issue_id":"thoughtbox-9a2.3","depends_on_id":"thoughtbox-9a2.2","type":"blocks","created_at":"2026-02-05T15:24:55.087754-06:00","created_by":"glassBead"},{"issue_id":"thoughtbox-9a2.3","depends_on_id":"thoughtbox-9a2","type":"parent-child","created_at":"2026-02-05T15:24:13.731173-06:00","created_by":"glassBead"}]} {"id":"thoughtbox-9a2.4","title":"Phase 4: Validation","description":"Automated and manual testing of hypotheses.","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-05T15:24:13.787751-06:00","updated_at":"2026-02-05T21:01:36.781163-06:00","closed_at":"2026-02-05T21:01:36.781163-06:00","close_reason":"All 9 hub operations validated via Dockerized MCP (Streamable HTTP). register, whoami, create_workspace, join_workspace, create_problem, post_message, read_channel, list_workspaces, workspace_status all pass.","created_by":"glassBead","dependencies":[{"issue_id":"thoughtbox-9a2.4","depends_on_id":"thoughtbox-9a2","type":"parent-child","created_at":"2026-02-05T15:24:13.78832-06:00","created_by":"glassBead"},{"issue_id":"thoughtbox-9a2.4","depends_on_id":"thoughtbox-9a2.3","type":"blocks","created_at":"2026-02-05T15:24:55.13071-06:00","created_by":"glassBead"}]} {"id":"thoughtbox-9a2.5","title":"Phase 5: Decision (Accept/Reject)","description":"User confirms accept or reject based on validation.","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-05T15:24:13.844528-06:00","updated_at":"2026-02-05T21:01:42.674469-06:00","closed_at":"2026-02-05T21:01:42.674469-06:00","close_reason":"ACCEPT: Hub design validated. All hypotheses confirmed — progressive disclosure works, operations round-trip through Docker MCP, channel messaging functional.","created_by":"glassBead","dependencies":[{"issue_id":"thoughtbox-9a2.5","depends_on_id":"thoughtbox-9a2.4","type":"blocks","created_at":"2026-02-05T15:24:55.173831-06:00","created_by":"glassBead"},{"issue_id":"thoughtbox-9a2.5","depends_on_id":"thoughtbox-9a2","type":"parent-child","created_at":"2026-02-05T15:24:13.845104-06:00","created_by":"glassBead"}]} {"id":"thoughtbox-9x7","title":"Eval: dual Client instantiation — trace-listener and dataset-manager each create LangSmith Client","description":"Both LangSmithTraceListener and DatasetManager independently create LangSmith Client instances from config. Should share a single client instance. Defer to Phase 3 experiment runner. Review finding P2.","status":"closed","priority":3,"issue_type":"task","created_at":"2026-02-12T16:17:12.135509-06:00","updated_at":"2026-02-12T20:57:29.086828-06:00","closed_at":"2026-02-12T20:57:29.086828-06:00","close_reason":"Resolved by Phase 3 experiment runner: shared client factory eliminates dual instantiation, type alignment removes YAGNI types","created_by":"glassBead"} {"id":"thoughtbox-a4m","title":"Write gateway profile priming tests (red)","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-07T04:04:55.291694-06:00","updated_at":"2026-02-07T04:06:02.407365-06:00","closed_at":"2026-02-07T04:06:02.407365-06:00","close_reason":"6 tests written, 3 red as expected","created_by":"glassBead"} {"id":"thoughtbox-a5a","title":"M10: Run regression tests and build","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-06T02:27:53.508835-06:00","updated_at":"2026-02-06T02:33:25.330791-06:00","closed_at":"2026-02-06T02:33:25.330791-06:00","close_reason":"Closed","created_by":"glassBead"} {"id":"thoughtbox-aik","title":"Module 2: hub-event-callback","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-05T19:38:23.712278-06:00","updated_at":"2026-02-05T19:41:43.544957-06:00","closed_at":"2026-02-05T19:41:43.544957-06:00","close_reason":"Closed","created_by":"glassBead"} {"id":"thoughtbox-aqj","title":"Rename observatory tabs from Git metaphors to Thoughtbox primitives","status":"open","priority":2,"issue_type":"feature","created_at":"2026-02-09T12:52:41.648164-06:00","updated_at":"2026-02-09T12:52:41.648164-06:00","created_by":"glassBead"} {"id":"thoughtbox-au3","title":"Module 0: thought-store-adapter","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-05T19:38:23.593578-06:00","updated_at":"2026-02-05T19:39:15.059995-06:00","closed_at":"2026-02-05T19:39:15.059995-06:00","close_reason":"Closed","created_by":"glassBead"} {"id":"thoughtbox-b3q","title":"Wire Hub into MCP Transport","status":"closed","priority":2,"issue_type":"epic","created_at":"2026-02-05T19:35:22.242735-06:00","updated_at":"2026-02-05T19:50:23.162615-06:00","closed_at":"2026-02-05T19:50:23.162615-06:00","close_reason":"Closed","created_by":"glassBead"} {"id":"thoughtbox-b6c","title":"Fix: branch thoughts not surfaced in get_structure, read_thoughts, deep_analysis","status":"closed","priority":1,"issue_type":"bug","created_at":"2026-02-06T08:29:56.819618-06:00","updated_at":"2026-02-06T08:33:25.514113-06:00","closed_at":"2026-02-06T08:33:25.514113-06:00","close_reason":"Fixed: getAllThoughts() and getBranchIds() added to storage interface, gateway operations updated to use them","created_by":"glassBead"} {"id":"thoughtbox-bo6","title":"M10: Wire mcpSessionId through server-factory.ts","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-06T02:27:46.401025-06:00","updated_at":"2026-02-06T02:33:25.291614-06:00","closed_at":"2026-02-06T02:33:25.291614-06:00","close_reason":"Closed","created_by":"glassBead"} {"id":"thoughtbox-bz3","title":"M9: Runtime wiring + observational docs + direct-use tests","status":"closed","priority":1,"issue_type":"feature","created_at":"2026-02-06T00:28:29.861098-06:00","updated_at":"2026-02-06T00:34:27.583942-06:00","closed_at":"2026-02-06T00:34:27.583942-06:00","close_reason":"M9 implemented: env var wiring, extended cipher delivery, auto-branch, 8 tests, 3 docs, direct-use validated","created_by":"glassBead"} {"id":"thoughtbox-c69","title":"Fix tool description discoverability gaps across 3 description strings","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-07T18:26:13.636809-06:00","updated_at":"2026-02-07T18:28:40.441607-06:00","closed_at":"2026-02-07T18:28:40.441607-06:00","close_reason":"All 3 description strings updated and verified","created_by":"glassBead"} {"id":"thoughtbox-cf6","title":"Implement gateway profile priming (green)","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-07T04:04:58.285748-06:00","updated_at":"2026-02-07T04:07:49.08818-06:00","closed_at":"2026-02-07T04:07:49.08818-06:00","close_reason":"6/6 tests green, 291 total pass","created_by":"glassBead","dependencies":[{"issue_id":"thoughtbox-cf6","depends_on_id":"thoughtbox-a4m","type":"blocks","created_at":"2026-02-07T04:05:07.451491-06:00","created_by":"glassBead"}]} {"id":"thoughtbox-ct8","title":"Knowledge graph disabled in Docker: better-sqlite3 native bindings mismatch","description":"All 12 knowledge graph behavioral tests (KG-001 through KG-012) fail because knowledgeHandler is undefined. Root cause: better-sqlite3 native module was compiled for macOS (darwin) but the Thoughtbox server runs in a Linux Docker container. Docker logs confirm: 'Could not locate the bindings file.' The Dockerfile needs to rebuild better-sqlite3 inside the container (npm rebuild better-sqlite3) during the Docker build step. The node_modules/better-sqlite3/build/Release/better_sqlite3.node on the host is a macOS Mach-O binary, not a Linux ELF binary.","status":"closed","priority":1,"issue_type":"bug","created_at":"2026-02-08T05:50:49.041908-06:00","updated_at":"2026-02-08T05:55:45.697897-06:00","closed_at":"2026-02-08T05:55:45.697897-06:00","close_reason":"Fixed: added npm rebuild better-sqlite3 to Dockerfile production stage","created_by":"glassBead"} {"id":"thoughtbox-cye","title":"Port assumptions registry and wire assumption-tracker hook","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-13T00:39:00.270822-06:00","updated_at":"2026-02-13T00:41:49.832567-06:00","closed_at":"2026-02-13T00:41:49.832567-06:00","close_reason":"Code committed; hooks need manual deployment","created_by":"glassBead"} {"id":"thoughtbox-efp","title":"Module 1: hub-task-store","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-05T19:38:23.653124-06:00","updated_at":"2026-02-05T19:40:18.990142-06:00","closed_at":"2026-02-05T19:40:18.990142-06:00","close_reason":"Closed","created_by":"glassBead"} {"id":"thoughtbox-f51","title":"Implement LangSmith evaluation foundation infrastructure","status":"closed","priority":1,"issue_type":"feature","created_at":"2026-02-11T21:25:18.979391-06:00","updated_at":"2026-02-11T21:32:31.928782-06:00","closed_at":"2026-02-11T21:32:31.928782-06:00","close_reason":"Spec written and Phase 1 infrastructure implemented","created_by":"glassBead"} {"id":"thoughtbox-g3r","title":"Flatten nested sub-operation arg shapes for session/notebook/mental_models","description":"Currently, session/notebook/mental_models operations require double-nested args: { operation: 'session', args: { operation: 'list', args: { limit: 3 } } }. This is 'present-at-hand' (Heidegger) — agents must think about the gateway's internal routing structure. Goal: flatten to { operation: 'session.list', limit: 3 } or similar so operations feel 'ready-at-hand'. Affects gateway-handler.ts handleSession/handleNotebook/handleMentalModels methods (lines ~734-778). Contrast with thought/deep_analysis/knowledge which already use flat args.","status":"open","priority":2,"issue_type":"feature","created_at":"2026-02-07T01:58:50.414241-06:00","updated_at":"2026-02-07T01:59:00.063726-06:00","created_by":"glassBead"} {"id":"thoughtbox-g9q","title":"Eval: unbounded list accumulation in DatasetManager listDatasets/getDatasetExamples","description":"listDatasets() and getDatasetExamples() accumulate all results into arrays with no upper bound. Large datasets could cause OOM. Need to enforce max or stream results. Review finding P1.","status":"closed","priority":2,"issue_type":"bug","created_at":"2026-02-12T16:17:00.137148-06:00","updated_at":"2026-02-12T21:10:48.554914-06:00","closed_at":"2026-02-12T21:10:48.554914-06:00","close_reason":"Fixed in Phase 2 review cleanup","created_by":"glassBead"} {"id":"thoughtbox-ih7","title":"M11: Implement dependencies (GREEN)","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-06T03:15:45.934382-06:00","updated_at":"2026-02-06T03:20:33.674737-06:00","closed_at":"2026-02-06T03:20:33.674737-06:00","close_reason":"Closed","created_by":"glassBead","dependencies":[{"issue_id":"thoughtbox-ih7","depends_on_id":"thoughtbox-kdy","type":"blocks","created_at":"2026-02-06T03:15:54.047305-06:00","created_by":"glassBead"}]} {"id":"thoughtbox-ike","title":"Eval: setTimeout timers not cancelled on trace-listener detach()","description":"onSessionEnded sets a 5-second setTimeout for cleanup. If detach() is called before the timer fires, the callback runs against cleared state. Timer references should be tracked and cleared in detach(). Review finding P1.","status":"closed","priority":2,"issue_type":"bug","created_at":"2026-02-12T16:17:04.668739-06:00","updated_at":"2026-02-12T21:10:48.563716-06:00","closed_at":"2026-02-12T21:10:48.563716-06:00","close_reason":"Fixed in Phase 2 review cleanup","created_by":"glassBead"} {"id":"thoughtbox-kdy","title":"M11: Write dependency tests (RED)","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-06T03:15:43.536807-06:00","updated_at":"2026-02-06T03:18:55.893611-06:00","closed_at":"2026-02-06T03:18:55.893611-06:00","close_reason":"Closed","created_by":"glassBead","dependencies":[{"issue_id":"thoughtbox-kdy","depends_on_id":"thoughtbox-rgf","type":"blocks","created_at":"2026-02-06T03:15:53.985257-06:00","created_by":"glassBead"}]} {"id":"thoughtbox-lx6","title":"Observatory GitHub Lite UI Redesign","status":"closed","priority":1,"issue_type":"feature","created_at":"2026-02-08T03:09:13.125019-06:00","updated_at":"2026-02-08T04:20:57.342107-06:00","closed_at":"2026-02-08T04:20:57.342107-06:00","close_reason":"UI redesign verified visually in browser - GitHub Primer dark theme renders correctly","created_by":"glassBead"} {"id":"thoughtbox-oaa","title":"Operations catalogs for all handler domains","status":"closed","priority":1,"issue_type":"feature","created_at":"2026-02-08T10:03:26.956017-06:00","updated_at":"2026-02-08T10:12:33.433686-06:00","closed_at":"2026-02-08T10:12:33.433686-06:00","close_reason":"Operations catalogs created for all 5 remaining handler domains (gateway, init, session, knowledge, hub). Resources registered, ListResources updated, per-operation embedding added, stage errors enhanced, hub vocabulary added.","created_by":"glassBead"} {"id":"thoughtbox-pum","title":"Eval: no batch chunking for createExamples in DatasetManager","description":"addCollectionExamples and addDeploymentExamples pass all tasks in a single createExamples call with no chunking. Large task arrays could hit API payload limits. Review finding P1.","status":"closed","priority":2,"issue_type":"bug","created_at":"2026-02-12T16:17:02.2135-06:00","updated_at":"2026-02-12T21:10:48.560113-06:00","closed_at":"2026-02-12T21:10:48.560113-06:00","close_reason":"Fixed in Phase 2 review cleanup","created_by":"glassBead"} {"id":"thoughtbox-q4j","title":"M10: Harden hub-tool-handler.ts with per-session identity map","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-06T02:27:44.09956-06:00","updated_at":"2026-02-06T02:33:25.278312-06:00","closed_at":"2026-02-06T02:33:25.278312-06:00","close_reason":"Closed","created_by":"glassBead"} {"id":"thoughtbox-rgf","title":"M11: Extend hub types for dependencies","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-06T03:15:41.155762-06:00","updated_at":"2026-02-06T03:17:24.495606-06:00","closed_at":"2026-02-06T03:17:24.495606-06:00","close_reason":"Closed","created_by":"glassBead"} {"id":"thoughtbox-rkv","title":"Fix hub-handler event payloads — data shape mismatch causes ? in observatory","status":"open","priority":1,"issue_type":"bug","created_at":"2026-02-09T12:52:39.735562-06:00","updated_at":"2026-02-09T12:52:39.735562-06:00","created_by":"glassBead"} {"id":"thoughtbox-t6c","title":"Module 6: wiring-integration","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-05T19:38:23.94746-06:00","updated_at":"2026-02-05T19:50:05.744276-06:00","closed_at":"2026-02-05T19:50:05.744276-06:00","close_reason":"Closed","created_by":"glassBead"} {"id":"thoughtbox-twu","title":"Gateway stage never resets for sub-agent MCP sessions","description":"Sub-agents spawned via Claude Code Task tool share the parent's MCP connection rather than getting their own fresh connection. This means sub-agent sessions start at stage_3 (fully initialized) instead of Stage 0 (entry). Progressive disclosure tests GW-010 and GW-011 are untestable because cipher/thought operations succeed at all stages. Init handler list_sessions filters (limit, search, project) are also silently ignored (GW-002, GW-003) — this may be a separate bug in the init handler, but was discovered in the same test context. The stage enforcement code exists in gateway-handler.ts (createStageError at line 397-411) but is unreachable when sessions start pre-initialized.","status":"closed","priority":2,"issue_type":"bug","created_at":"2026-02-08T05:50:55.208533-06:00","updated_at":"2026-02-09T12:38:24.507162-06:00","closed_at":"2026-02-09T12:38:24.507162-06:00","close_reason":"Fixed in commit 625d1fa. Per-session stage tracking via sessionStages Map in GatewayHandler. 7 new tests. Independently verified by judge agent.","created_by":"glassBead"} {"id":"thoughtbox-u1t","title":"Enhance session_start.sh and pre_compact.sh for handoff","status":"closed","priority":1,"issue_type":"task","created_at":"2026-02-13T00:38:58.941208-06:00","updated_at":"2026-02-13T00:41:49.814803-06:00","closed_at":"2026-02-13T00:41:49.814803-06:00","close_reason":"Code committed; hooks need manual deployment","created_by":"glassBead"} {"id":"thoughtbox-vxw","title":"Research: MCP tool name duplication in sub-agent spawning","status":"closed","priority":1,"issue_type":"task","created_at":"2026-02-08T00:36:39.895591-06:00","updated_at":"2026-02-08T00:50:25.656471-06:00","closed_at":"2026-02-08T00:50:25.656471-06:00","close_reason":"Research complete. Mechanism identified: mcpServers in agent frontmatter + inherited parent tools = duplicate tool names. Fix: remove mcpServers, use ToolSearch. Known Claude Code bug (GH #10668, #10704, #21560).","created_by":"glassBead"} {"id":"thoughtbox-w8u","title":"Write SPEC-EVAL-001 unified evaluation system specification","status":"closed","priority":1,"issue_type":"task","created_at":"2026-02-11T21:25:17.95623-06:00","updated_at":"2026-02-11T21:32:31.909597-06:00","closed_at":"2026-02-11T21:32:31.909597-06:00","close_reason":"Spec written and Phase 1 infrastructure implemented","created_by":"glassBead"} {"id":"thoughtbox-wcs","title":"Open-Ended Workflow Discovery via Thoughtbox","status":"closed","priority":2,"issue_type":"feature","created_at":"2026-02-06T07:59:58.454614-06:00","updated_at":"2026-02-06T08:15:49.650461-06:00","closed_at":"2026-02-06T08:15:49.650461-06:00","close_reason":"135-thought session completed. Key findings: friction hierarchy predicts agent behavior, branching has retrieval gap bug, three design principles (friction gradient, visible feedback, composition) identified, 8 new workflow classes proposed, concrete implementation roadmap produced.","created_by":"glassBead"} {"id":"thoughtbox-wf7","title":"Eval: remove dead EvaluatorResult type and ~85 LOC YAGNI types from types.ts","description":"types.ts contains EvaluatorResult (zero consumers), MonitoringAlert, AlertSeverity, AlertType, and other Phase 4/5 forward-declared types that are dead code. Remove to reduce confusion. Review finding P2.","status":"closed","priority":3,"issue_type":"task","created_at":"2026-02-12T16:17:09.976776-06:00","updated_at":"2026-02-12T20:57:29.095522-06:00","closed_at":"2026-02-12T20:57:29.095522-06:00","close_reason":"Resolved by Phase 3 experiment runner: shared client factory eliminates dual instantiation, type alignment removes YAGNI types","created_by":"glassBead"} {"id":"thoughtbox-whf","title":"Implement capture-handoff.mjs","status":"closed","priority":1,"issue_type":"task","created_at":"2026-02-13T00:38:57.517525-06:00","updated_at":"2026-02-13T00:39:54.706445-06:00","closed_at":"2026-02-13T00:39:54.706445-06:00","close_reason":"capture-handoff.mjs implemented and tested","created_by":"glassBead"} {"id":"thoughtbox-wrq","title":"Add missing activity events — agent registration, reviews, claims, dependencies","status":"open","priority":2,"issue_type":"feature","created_at":"2026-02-09T12:52:45.518833-06:00","updated_at":"2026-02-09T12:52:45.518833-06:00","created_by":"glassBead"} {"id":"thoughtbox-xsr","title":"Eval: no circuit breaker in trace-listener safeAsync — repeated failures flood logs","description":"safeAsync catches errors but keeps retrying on every event. If LangSmith is down, every thought event produces a warning. Need a circuit breaker or exponential backoff to suppress repeated failures. Review finding P2.","status":"closed","priority":3,"issue_type":"task","created_at":"2026-02-12T16:17:14.53231-06:00","updated_at":"2026-02-12T21:10:48.57051-06:00","closed_at":"2026-02-12T21:10:48.57051-06:00","close_reason":"Fixed in Phase 2 review cleanup","created_by":"glassBead"}