Randall Stillwell 6ccb060bea feat(second-brain): open self-hosted Obsidian for crews — harness-agnostic v0.1

- MCP create-backlog-item for natural language Plan/Epic/Task authoring
- Worker non-code: optional worktree + per-harness allowlist + doc_review→published
- Harness registry: harness_providers/role_harness_routes, WorkerAdapter (minimal|crewai|langgraph|remote-ag-ui), AG-UI POST /api/ag-ui, Settings → Harnesses with Test
- Plans tree: backlog.listTree + createPlan/Epic/Task + Start→Convoy, web CRUD without CLI
- Fleet: /fleet rollup across workspaces + SSE /api/fleet/events

Decision: docs/decisions/2026-08-25-second-brain-os-pivot.md
Convoy: .convoys/second-brain-os-v01

Co-authored-by: orchestrator

2026-08-25 16:20:53 -05:00

2 KiB

Raw Blame History

kind	slug	title	plan_slug	epic_slug	status	priority	tenant_id	owner	cursor_todo_id	updated_at
task	worker-checkpoint-resume	Worker checkpoint and crash recovery	automation-foundation	apps-worker-v1	done	P1	global	unassigned	null	2026-08-12

Task summary

Persist worker session progress at configurable checkpoints so a crashed or stalled job resumes from the last checkpoint instead of restarting the brief from scratch.

Description

Benchmark: Autensa mission-control checkpoint & crash recovery — study their checkpoint restore API and per-task history UX; implement Echodo-native equivalents on Postgres.

Schema

Add worker_checkpoints (or JSONB on agent_runs) workspace-scoped:

agent_run_id, convoy_id, brief_id
checkpoint_seq, tool_history (bounded), files_touched, branch, worktree_path
occurred_at

Worker behavior

Save checkpoint after each tool round (configurable via WORKER_CHECKPOINT_EVERY_N_STEPS, default 1).
On job start: if open run has latest checkpoint, resume tool loop from saved state.
Orchestrator stall detector can trigger re-dispatch with resume_from_checkpoint: true.

Operator surface (minimal v1)

Mission control detail panel: checkpoint count + last checkpoint time (full history deferred).

Dependencies

Task-worker-minimal-tool-loop (baseline loop must exist).

Subtasks

Zod schemas in packages/shared/worker
Drizzle migration + workspace-scoped queries
Checkpoint write in minimal-session / handle-job
Resume path on re-dispatch
Env docs in docs/worker.md

Estimation

Acceptance criteria

Kill worker mid-brief; re-dispatch resumes from last tool step without re-cloning
Checkpoints filtered by workspace_id
Checkpoint payload does not store API keys or full file contents (paths + hashes only)

2 KiB Raw Blame History