Commit Graph

  • f93c1f0672 Reproduce the Colima timer fix: remove the 5-minute launchd StartInterval so com.joel.colima only starts the VM at boot instead of hammering colima start against a running instance Joel Hooks 2026-04-16 09:47:27 -07:00
  • 3394fa5b95 Reproduce the fast-but-earned healer escalation fix: add a rapid confirmation window and early host-path bailouts so real Colima collapse recovers faster without restoring single-tick panic cycles Joel Hooks 2026-04-16 08:40:29 -07:00
  • 4fcc853465 Reproduce the healer guardrail fix: persist Colima unhealthy streaks and cooldowns so k8s-reboot-heal stops single-tick warmup from force-cycling the VM, and document the new escalation contract Joel Hooks 2026-04-16 08:34:54 -07:00
  • 2f558b82e4 Reproduce the reboot-heal state fix: persist flannel and recovery markers across launchd ticks so Typesense is not re-warmed by repeated flannel restarts Joel Hooks 2026-04-15 17:57:14 -07:00
  • fca46a276d Reproduce the stale-tunnel reboot fix: make colima-tunnel restart on SSH port drift, harden k8s-reboot-heal against false Colima cycling, and make gateway status require real daemon reachability Joel Hooks 2026-04-15 17:37:35 -07:00
  • 4fa2bcee59 Reproduce the final reboot tunnel fix: stop forwarding 6443 from colima-tunnel and document that Caddy owns the port Joel Hooks 2026-04-14 17:49:03 -07:00
  • f96157b683 Reproduce the reboot root-cause fix: add a repo-canonical colima tunnel launchd service, wire it into the critical installer, and make gateway CLI inspect system launchd services Joel Hooks 2026-04-14 17:00:47 -07:00
  • 5ff6a9a7df Correct the boot-safe agent-mail runtime so launchd resolves joelhooks/mcp_agent_mail instead of hardcoding the legacy Dicklesworthstone checkout path Joel Hooks 2026-04-12 09:03:28 -07:00
  • 3384b45d9a Replace the failed ADR-0239 bridge with boot-safe LaunchDaemons for critical host services and retire user/501 bootstrap Joel Hooks 2026-04-12 08:57:27 -07:00
  • ca7a113b64 Harden Langfuse secret fallback so secrets lease error JSON cannot become base URLs or credentials Joel Hooks 2026-04-12 08:38:23 -07:00
  • 5fb084569f Install a headless boot bridge for critical launchd services and track gateway/typesense/agent-mail plists in repo Joel Hooks 2026-04-12 08:27:08 -07:00
  • ef5728a40c Recover reboot health checks: ignore terminal pods in joelclaw status and document Colima 8/16 plus host-worker 3111 invariants Joel Hooks 2026-04-12 08:11:22 -07:00
  • 8e40af7698 chore: auto-commit from o11y triage Joel Hooks 2026-04-12 04:15:55 -07:00
  • 56913ee882 Implement ADR-0238 by keeping machine review fallbacks out of Joel-facing Todoist projects and limiting tasks/triage to human-facing task surfaces. Joel Hooks 2026-04-08 15:48:52 -07:00
  • 1dc22d11dd Document ADR-0238 task-surface rules in the task-management skill so interactive Todoist capture keeps machine bookkeeping out of Joel-facing views. Joel Hooks 2026-04-07 22:19:33 -07:00
  • 3722ccaec6 Suppress low-signal recovered gateway relay events, add a regression test, and update the gateway docs. Joel Hooks 2026-04-07 21:11:29 -07:00
  • 5e1c26205f Bypass strict Front inbox pagination decoding too so Email Nag survives _pagination.next null on both inbox and conversation listing Joel Hooks 2026-04-06 15:22:51 -07:00
  • 651e563b58 Treat Front unread queries as unreplied and bypass strict pagination decoding so Email Nag stops failing on _pagination.next null Joel Hooks 2026-04-06 15:20:12 -07:00
  • bef5f63b74 Upsert the full channel_messages document during classification so Typesense keeps required timestamp sorting data and stops 400ing Joel Hooks 2026-04-06 15:09:06 -07:00
  • 51713a014e Raise joelclaw run and event detail GraphQL defaults because live detail queries are taking more than a minute under load Joel Hooks 2026-04-06 11:59:40 -07:00
  • 855d79ed4b Raise joelclaw runs GraphQL defaults because filtered self-hosted runs queries are still taking ~50s in practice Joel Hooks 2026-04-06 11:55:18 -07:00
  • 2c5d70d1da Fetch failed step outputs in parallel so joelclaw run no longer times out walking large failed traces Joel Hooks 2026-04-06 11:46:32 -07:00
  • cfd17eebfa Give gateway fallback time to save the session, raise joelclaw run/event detail GraphQL budgets, and stop self-healing investigator failed-run scans timing out Joel Hooks 2026-04-06 11:40:12 -07:00
  • b75428b0bf Degrade malformed channel-message-classify output to heuristic fallback instead of failing the workflow Joel Hooks 2026-04-06 09:29:36 -07:00
  • 2ab65429f4 Skip immutable Typesense id when patching channel_messages schema so channel-message-ingest stops flooding false failures Joel Hooks 2026-04-06 09:18:12 -07:00
  • 76b3b98df9 Probe /webhooks via endpoint fallback, switch Agent Secrets health to secrets status, and give joelclaw runs adaptive GraphQL timeouts Joel Hooks 2026-04-06 08:48:00 -07:00
  • 5022222c6b fix(docs-ingest): add NAS path fallback and incoming/ cleanup Joel Hooks 2026-04-06 08:07:30 -07:00
  • f9a4843bd8 talon: --no-extensions but keep sessions and skills for crisis agent Joel Hooks 2026-04-01 10:34:11 -07:00
  • 0c4ac7df57 talon: upgrade agent crisis recovery to full pi system agent sessions Joel Hooks 2026-04-01 10:33:26 -07:00
  • af0e5319e3 Add capture_progress skill so agents turn 'capture this' into durable truth in the right artifact Joel Hooks 2026-03-28 11:44:18 -07:00
  • 9ee695f6e7 Remove the injected non-trivial-task ack/pause reminder from session-lifecycle turn prompts Joel Hooks 2026-03-28 11:14:32 -07:00
  • 806352471d Fix joelclaw log capability provenance handling and replace deprecated Effect formatter usage across SDK adapters Joel Hooks 2026-03-28 10:57:53 -07:00
  • 825972ced3 Switch the gateway context to conversation_threads for project momentum and relationship tracking Joel Hooks 2026-03-28 09:28:23 -07:00
  • babf57dd4e Implement thread-oriented conversation intelligence for channel messages, thread aggregation, and vault-aware enrichment Joel Hooks 2026-03-28 09:11:56 -07:00
  • 4832974cf4 ADR-0236: Wire realtime Slack + Front → Typesense indexing — Slack messages emit channel/message.received via Inngest HTTP API on each message (both invoke and passive intel paths). Front inbound emails emit channel/message.received from front-notify.ts index-channel-message step. Added 'email' to channel type unions in client.ts and channel-message-ingest.ts. Existing channel-message-ingest → channel-message-classify pipeline handles indexing and classification. Joel Hooks 2026-03-28 08:38:17 -07:00
  • c29c9f753e ADR-0235: Add business intelligence to gateway context — query Typesense for recent Slack messages (7d, grouped by channel) and email threads on operator interaction. Lease typesense_api_key at startup via agent-secrets. Gateway now answers 'what's important' with actual Slack conversations and email threads, not just infra metrics. Joel Hooks 2026-03-28 08:17:41 -07:00
  • 3f83571dee ADR-0235: Fix context gathering speed — replace joelclaw runs (15s timeout) with joelclaw status (570ms). Drop old assistant snippets from session overflow recovery in daemon.ts. Total context gather now ~600ms instead of 8s timeout. Joel Hooks 2026-03-28 07:24:18 -07:00
  • a51e2b822d ADR-0235: Strip old assistant responses from context overflow recovery — stale triage summaries (NAS/CWA/security advice) were being injected into fresh gateway sessions via buildCompressionSummary(), causing the gateway to regurgitate old strategic advice instead of answering from live system data. Keep recent user messages for continuity, drop assistant snippets entirely. Joel Hooks 2026-03-28 07:21:33 -07:00
  • ba7a1a8487 ADR-0235: Add on-demand system context gathering for gateway — when operator sends a message, gather slog (recent 15 entries), OTEL stats (24h error rate), recent runs (last 5), and failed runs (24h grouped by function) in parallel via CLI calls. Injected alongside event buffer into system prompt so gateway can answer 'what's important right now' and 'summarize this week' with real data. Joel Hooks 2026-03-27 19:56:10 -07:00
  • 0bb263f2bd ADR-0235 cleanup: remove dead code from gateway extension — isDigestOnlyMode, logAutomationCounter, shouldInjectHeartbeat, isHeartbeatDegraded, buildQuietDigestSummary, buildSubscriptionMessage, DrainOptions interface, AUTO_MESSAGE_DIGEST_ONLY_THRESHOLD, HEARTBEAT_FULL_CHECK_MS, automatedInjectedCount, lastFullHeartbeatCheckTs. -114 lines. Joel Hooks 2026-03-27 19:39:02 -07:00
  • 00775a73e7 ADR-0235: Switch gateway to demand-driven events — buffer system events silently in Redis context buffer, only inject critical events (errors/failures/degraded) as LLM messages, prepend accumulated context to system prompt when operator sends a message. Removes broadcast pattern where every drain cycle forced an LLM response. Context buffer: max 50 events, 24h TTL, grouped by type, flushed on operator interaction. Joel Hooks 2026-03-27 19:26:53 -07:00
  • 9bd62e6093 feat: PDS dual-write for slog and memory observations Joel Hooks 2026-03-21 21:25:01 -07:00
  • 1c7533970d Fix packages/system-bus type errors by tightening queue observer/test mocks, adding missing event schemas, and guarding optional unions Joel Hooks 2026-03-21 21:05:38 -07:00
  • 173f0a310e health.sh: fix PDS port to use Docker mapping (9627) not port-forward (2583) Joel Hooks 2026-03-21 20:56:27 -07:00
  • c83af73d2e k8s skill: document kubeconfig port drift pattern and self-heal fix Joel Hooks 2026-03-21 20:52:59 -07:00
  • ba8ff821c9 health.sh: add kubeconfig self-heal on port drift + fix stale test false positive Joel Hooks 2026-03-21 20:51:02 -07:00
  • a92e88ab11 feat: add nas-hdd PV/PVC for HDD RAID5 tier (50TB, /volume1/joelclaw) — k8s pods can now access books, docs-artifacts, backups, archives over 10GbE NFS. Update k8s + system-architecture skills with both NAS tiers. Joel Hooks 2026-03-19 21:07:08 -07:00
  • d754e99dac docs: capture k8s↔NAS NFS networking lore in k8s skill, system-architecture skill, and architecture docs — route via col0, NFS PVs, access patterns, verification steps (ADR-0088 Phase 2.5) Joel Hooks 2026-03-19 20:25:41 -07:00
  • d89c2a6478 feat: solve k8s↔NAS networking — add LAN route via col0 bridge, create NFS PV/PVC for NAS NVMe tier, persist route in Colima provision + tunnel scripts (ADR-0088 Phase 2.5) Joel Hooks 2026-03-19 19:41:50 -07:00
  • 53d5576ab2 fix: update docs-api k8s manifest to v0.2.0 image tag (20260319-192432), make docs status CLI gracefully handle missing v1 docs_chunks collection (dropped to reclaim 79G disk space from Typesense snapshots + superseded collection) Joel Hooks 2026-03-19 19:31:20 -07:00
  • cc600aaf32 fix(docs-api): remove embedding from query_by — v2 has no auto-embed field Joel Hooks 2026-03-18 22:27:47 -07:00
  • aa87a5b009 skills(k8s): document stale SSH mux socket fix after Colima restart Joel Hooks 2026-03-18 22:24:31 -07:00
  • c84d808d10 fix(docs-api): graceful degradation + unhandled rejection handler Joel Hooks 2026-03-18 21:05:54 -07:00
  • 5dac05f19f fix: truncate to 8K chars for embed — some retrieval_text still exceeding nomic context Joel Hooks 2026-03-18 09:30:17 -07:00
  • 1445ace3a2 fix: lower embed truncation to 16K chars (nomic 8192-token limit) Joel Hooks 2026-03-18 07:18:20 -07:00
  • 44fa09abb5 fix: singleton mode cancel (not skip) so stale batch runs can be replaced Joel Hooks 2026-03-18 06:22:42 -07:00
  • 78ebf41148 fix: disable vector_query for v2 collection (raw float[], no auto-embed) Joel Hooks 2026-03-17 22:27:57 -07:00
  • eabc200a47 fix: truncate text to 28K chars before ollama embed (nomic 8192-token limit) Joel Hooks 2026-03-17 22:23:39 -07:00
  • 90c74d4f15 feat: GPU-accelerated ollama embeddings for docs pipeline v2 (ADR-0234) Joel Hooks 2026-03-17 22:19:44 -07:00
  • e343569460 fix(docs-reindex-v2): batch Typesense import in groups of 100 Joel Hooks 2026-03-17 22:13:40 -07:00
  • 8d1411c601 Add ADR-0234 docs pipeline v2 CLI commands and status/help updates? Joel Hooks 2026-03-17 22:12:09 -07:00
  • cb494d88f1 feat: chain docs/ingest.completed → docs/reindex-v2.requested (ADR-0234) Joel Hooks 2026-03-17 22:10:19 -07:00
  • 61246f574b skills: refresh pdf-brain and pdf-brain-ingest for ADR-0234 pipeline v2 Joel Hooks 2026-03-17 22:08:43 -07:00
  • 4a6700d506 Upgrade docs-api for ADR-0234 by switching to a configurable chunks collection, adding NAS artifact endpoints, and enriching agentic HATEOAS search context Joel Hooks 2026-03-17 22:08:24 -07:00
  • d8719c89d9 fix: add emitOtelEvent import to docs-reindex-batch Joel Hooks 2026-03-17 20:49:22 -07:00
  • ee8763b49d fix(docs-reindex-batch): step.sendEvent must be top-level, not nested in step.run Joel Hooks 2026-03-17 20:48:55 -07:00
  • 34122ceefd fix(docs-artifacts): NAS default + atomic writes for durability (ADR-0234) Joel Hooks 2026-03-17 20:04:30 -07:00
  • f117c0c0e7 Implement ADR-0234 docs pipeline v2 with artifact helpers, markdown-native chunking, nomic docs_chunks_v2, and reindex orchestration Joel Hooks 2026-03-17 20:02:43 -07:00
  • 84a4451cd3 ops: add Java to worker PATH for opendataloader-pdf Joel Hooks 2026-03-17 19:20:35 -07:00
  • ae620ad7d1 feat(docs-ingest): upgrade PDF extraction to opendataloader-pdf Joel Hooks 2026-03-17 19:19:32 -07:00
  • 3e4653bab7 Implement ADR-0233 phase 4 unified observability query CLI with otel session/system filters and o11y session/system correlation Joel Hooks 2026-03-17 16:04:45 -07:00
  • f8971ec24b x-api skill: add X Articles section — use agent-browser for long-form posts since v2 API returns 500 on article content. Document fallback pattern: API tweet lookup → detect article.title → agent-browser snapshot → discovery capture. Joel Hooks 2026-03-17 15:26:52 -07:00
  • 8edd9dd14d fix: PATCH sessionId+systemId onto existing otel_events collection (ADR-0233) Joel Hooks 2026-03-17 15:22:53 -07:00
  • afc2aed3c8 fix(sdk): include sessionId+systemId in OTEL emit payload (ADR-0233) Joel Hooks 2026-03-17 13:34:51 -07:00
  • 55e70bdec5 feat: add sessionId+systemId provenance to OTEL events (ADR-0233 Phase 3) Joel Hooks 2026-03-17 13:25:55 -07:00
  • 2b98f7c5b4 feat(system-bus): add sessionId+systemId provenance to slog Typesense sync + system-logger (ADR-0233 Phase 2) Joel Hooks 2026-03-17 13:10:00 -07:00
  • cd529fd54f docs: update TOOLS.md + workflow-rig skill with DAG rig capabilities Joel Hooks 2026-03-17 13:04:13 -07:00
  • e206be5710 clean: remove stale files, pause ADR-0230, verify DAG pipeline end-to-end Joel Hooks 2026-03-17 12:23:33 -07:00
  • 1354556514 evidence: autonomous commit from restate-worker k8s pod (direct exec) joelclaw-agent 2026-03-17 19:22:16 +00:00
  • 720b4fd931 Document the 2026-03-17 Talon Colima/Restate stability probes Joel Hooks 2026-03-17 11:02:42 -07:00
  • c1ed515d19 docs: update k8s, system-architecture, talon skills with Colima stability findings Joel Hooks 2026-03-17 11:01:42 -07:00
  • 30b52fed5b fix: make /dev/kvm mount optional in restate-worker manifest Joel Hooks 2026-03-17 10:53:21 -07:00
  • ee04c6fc8a chore: update restate-worker image to 20260317-093706 with retry caps Joel Hooks 2026-03-17 10:08:47 -07:00
  • a298aaf926 fix: add retry caps to dagWorker (5) and dagOrchestrator (3) + purge Restate state Joel Hooks 2026-03-17 09:35:36 -07:00
  • 267659df11 chore: update restate-worker image tag to 20260317-091052 Joel Hooks 2026-03-17 09:31:15 -07:00
  • ee4e098b0b fix: restate-worker Dockerfile adds e2fsprogs, manifest uses Secret for pi-auth Joel Hooks 2026-03-17 08:59:23 -07:00
  • cf55d75613 feat: one-shot microVM exec model — close the host↔guest gap Joel Hooks 2026-03-17 08:32:09 -07:00
  • 70868f9fd1 docs: add slog provenance usage to SYSTEM.md (ADR-0233) Joel Hooks 2026-03-17 08:29:20 -07:00
  • 9e6ba1e895 feat(sdk): add sessionId + systemId to slog-cli adapter (ADR-0233) Joel Hooks 2026-03-17 08:28:34 -07:00
  • 957929e510 feat: guest-runner.sh + rootfs init wiring for microVM exec protocol Joel Hooks 2026-03-17 08:24:29 -07:00
  • 1a4e9d52c1 prd: Firecracker microVM exec gap closure — 6 stories to prove end-to-end Joel Hooks 2026-03-17 08:23:05 -07:00
  • 4679c384ec audit: fresh-eyes analysis of Firecracker microVM execution path Joel Hooks 2026-03-17 08:14:51 -07:00
  • 9f9f4c49e5 Publish DAG completion to Redis pub/sub for pi session extension notifications Joel Hooks 2026-03-17 07:51:04 -07:00
  • c58fc8ec20 Refresh workflow-rig skill and workloads docs so they describe the proven 2026-03-17 runtime state Joel Hooks 2026-03-17 07:48:18 -07:00
  • 51e2ed7af2 Wire worker-heartbeat pi extension: proper ExtensionAPI format, symlinked in Dockerfile Joel Hooks 2026-03-17 07:30:28 -07:00
  • 5f5af27212 Worker improvements: pre-cloned repo cache, heartbeat extension, workspace helper Joel Hooks 2026-03-17 07:26:48 -07:00
  • 8ef1329760 feat: LRU caching for docs, chunks, TOC (ADR-0228 Phase 5) — autonomous agent joelclaw-agent 2026-03-17 14:26:22 +00:00
  • 59b8bf0746 Fix auth staleness + timeout: hostPath mount for pi auth, bump Restate timeouts Joel Hooks 2026-03-17 07:24:09 -07:00
  • 2b837b4dbf feat: TTL cache for doc lookups (ADR-0228 Phase 5) — autonomous joelclaw-agent 2026-03-17 14:22:31 +00:00
  • 3d8a46a083 feat: add LRU caching for doc/chunk/TOC lookups (ADR-0228 Phase 5) joelclaw-agent 2026-03-17 14:20:04 +00:00