docs(plans): add Autensa benchmark tasks to automation foundation roadmap

Map reliability, operator UX, and spend-control follow-ons onto epics 1–9 so worker, webhooks, and fleet work have bead-scale task briefs. Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-03 10:48:36 -05:00 · 2026-08-03 10:48:36 -05:00 · 35def387ae
commit 35def387ae
parent 2f6adf4382
20 changed files with 886 additions and 34 deletions
--- a/plans/Plan-automation-foundation/Epic-apps-orchestrator/Epic-apps-orchestrator.md
+++ b/plans/Plan-automation-foundation/Epic-apps-orchestrator/Epic-apps-orchestrator.md
@ -7,7 +7,7 @@ status: done
 priority: P1
 tenant_id: global
 cursor_epic_id: null
-updated_at: "2026-07-03"
+updated_at: "2026-08-03"
 ---

 # Epic objective
@ -55,6 +55,12 @@ Each job calls `resolveModel(workspace, roleId, convoy, brief)` and passes `Reso
 - [x] [Task-shared-orchestrator-driver](./Task-shared-orchestrator-driver.md)
 - [x] [Task-apps-orchestrator-service](./Task-apps-orchestrator-service.md)

+## Follow-on tasks (Autensa benchmark)
+
+| Task | Link |
+|------|------|
+| Agent health auto-nudge & zombie recovery | [Task-orchestrator-auto-nudge](./Task-orchestrator-auto-nudge.md) |
+
 ## Note

 Supersedes [Epic-optional-orchestrator](../../Plan-agent-coordination/Epic-optional-orchestrator/Epic-optional-orchestrator.md) deferred stub with a concrete Echodo-native design.
--- a/plans/Plan-automation-foundation/Epic-apps-orchestrator/Task-orchestrator-auto-nudge.md
+++ b/plans/Plan-automation-foundation/Epic-apps-orchestrator/Task-orchestrator-auto-nudge.md
@ -0,0 +1,64 @@
+---
+kind: task
+slug: orchestrator-auto-nudge
+title: Orchestrator agent health — auto-nudge and zombie recovery
+plan_slug: automation-foundation
+epic_slug: apps-orchestrator
+status: draft
+priority: P2
+tenant_id: global
+owner: unassigned
+cursor_todo_id: null
+updated_at: "2026-08-03"
+---
+
+# Task summary
+
+Extend stall detection beyond marking convoys `blocked`: detect zombie agent runs, optionally re-dispatch or nudge, and emit health events for mission control / fleet feed.
+
+## Description
+
+**Benchmark:** Autensa agent health monitoring with auto-nudge and convoy health panel. Echodo already has `ORCHESTRATOR_STALL_MINUTES` and `detectStalledConvoys` — this task adds **recovery actions** and observability.
+
+### Health states
+
+| State | Trigger | Default action |
+|-------|---------|----------------|
+| **Stalled** | No heartbeat on open `agent_run` for N minutes | Mark convoy `blocked`; surface in gate inbox sidebar |
+| **Zombie** | Run open but worker process gone | Re-enqueue job with `resume_from_checkpoint` when Epic 7 checkpoint task ships |
+| **Nudge** | Soft stall (configurable) | Log `convoy_events` health event; optional Redis ping to worker |
+
+### Config (env)
+
+- `ORCHESTRATOR_NUDGE_MINUTES` (optional, < stall threshold)
+- `ORCHESTRATOR_AUTO_REDISPATCH` (default false — opt-in for unattended dogfood)
+
+### Events
+
+Append `convoy_events` with `event_type: health` and metadata `{ reason, action, agent_run_id }` for activity feed (Epic 9).
+
+## Dependencies
+
+- Epic 6 orchestrator service (shipped)
+- [Task-worker-checkpoint-resume](../Epic-apps-worker-v1/Task-worker-checkpoint-resume.md) for meaningful re-dispatch
+
+## Subtasks
+
+- [ ] Extend `detectStalledConvoys` with nudge vs hard-stall tiers
+- [ ] Optional re-dispatch path on zombie detection
+- [ ] `convoy_events` health event writer
+- [ ] Document in `docs/orchestrator.md`
+
+## Estimation
+
+M
+
+## Acceptance criteria
+
+- [ ] Stalled run moves convoy to `blocked` and writes health event
+- [ ] With `ORCHESTRATOR_AUTO_REDISPATCH=1`, zombie run triggers new job enqueue
+- [ ] No unbounded re-dispatch loops (max retries per brief configurable)
+
+## Links
+
+- Epic: `./Epic-apps-orchestrator.md`
--- a/plans/Plan-automation-foundation/Epic-apps-worker-v1/Epic-apps-worker-v1.md
+++ b/plans/Plan-automation-foundation/Epic-apps-worker-v1/Epic-apps-worker-v1.md
@ -7,7 +7,7 @@ status: in_progress
 priority: P1
 tenant_id: global
 cursor_epic_id: null
-updated_at: "2026-07-03"
+updated_at: "2026-08-03"
 ---

 # Epic objective
@ -25,11 +25,13 @@ Role prompt = `.cursor/agents/role-*.md` + brief path. Model endpoint points at
 - `WorkerSession` adapter interface (for future optional OpenHands plug-in).
 - `pnpm lint && pnpm test` in allowlisted commands.
 - Branch + PR draft output (forge adapter from Epic 3).
+- Checkpoint/resume and operator mid-build notes (Autensa benchmark — see tasks below).

 **Out of scope**

 - Browser automation, multi-agent delegation.
 - Feature parity with OpenHands.
+- OpenClaw Gateway coupling.

 ## Dependencies

@ -43,7 +45,12 @@ Role prompt = `.cursor/agents/role-*.md` + brief path. Model endpoint points at

 - [ ] One brief from `.convoys/` completes unattended on connected repo.
 - [ ] `complete_task` includes model alias + token totals.
+- [ ] Crashed worker can resume from checkpoint (follow-on task).

 ## Tasks

-_To be decomposed when epic graduates to `ready`._
+| Task | Link | Status |
+|------|------|--------|
+| Minimal tool loop + worktree | [Task-worker-minimal-tool-loop](./Task-worker-minimal-tool-loop.md) | ready |
+| Checkpoint & crash recovery | [Task-worker-checkpoint-resume](./Task-worker-checkpoint-resume.md) | draft |
+| Operator notes mid-build | [Task-worker-operator-notes](./Task-worker-operator-notes.md) | draft |
--- a/plans/Plan-automation-foundation/Epic-apps-worker-v1/Task-worker-checkpoint-resume.md
+++ b/plans/Plan-automation-foundation/Epic-apps-worker-v1/Task-worker-checkpoint-resume.md
@ -0,0 +1,66 @@
+---
+kind: task
+slug: worker-checkpoint-resume
+title: Worker checkpoint and crash recovery
+plan_slug: automation-foundation
+epic_slug: apps-worker-v1
+status: draft
+priority: P1
+tenant_id: global
+owner: unassigned
+cursor_todo_id: null
+updated_at: "2026-08-03"
+---
+
+# 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](https://github.com/crshdn/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](./Task-worker-minimal-tool-loop.md) (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
+
+M
+
+## 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)
+
+## Links
+
+- Epic: `./Epic-apps-worker-v1.md`
+- Blocks: reliable unattended dogfood for Epic 8
--- a/plans/Plan-automation-foundation/Epic-apps-worker-v1/Task-worker-minimal-tool-loop.md
+++ b/plans/Plan-automation-foundation/Epic-apps-worker-v1/Task-worker-minimal-tool-loop.md
@ -0,0 +1,67 @@
+---
+kind: task
+slug: worker-minimal-tool-loop
+title: apps/worker — minimal tool loop and worktree-per-brief
+plan_slug: automation-foundation
+epic_slug: apps-worker-v1
+status: ready
+priority: P1
+tenant_id: global
+owner: unassigned
+cursor_todo_id: null
+updated_at: "2026-08-03"
+agent_prompt: |
+  Epic: apps/worker v1. Read docs/worker.md and docs/adr-model-routing.md.
+  Consume orchestrator jobs from Redis; resolve model via workspace registry.
+  claim_task / complete_task parity with MCP. Run pnpm lint && pnpm type-check
+  && pnpm test before done.
+---
+
+# Task summary
+
+Ship the baseline `apps/worker` consumer: dequeue `role_dispatch` jobs, run `WorkerSession` tool loop against a git worktree, and close with model + token metadata on `complete_task`.
+
+## Description
+
+Core v1 scope from [Epic-apps-worker-v1](./Epic-apps-worker-v1.md). Builds on existing package skeleton (`handle-job.ts`, `minimal-session.ts`, allowlisted tools, worktree helpers).
+
+### Worker loop
+
+- `BLPOP` on `echodo:orchestrator:jobs` (same key as orchestrator).
+- Load brief path + role from job payload; resolve model via shared resolver.
+- `claim_task` on matching backlog item; run `read_file` / `write_file` / `grep` / `run_command` (allowlist: `pnpm lint`, `pnpm type-check`, `pnpm test`, `git status`, `git diff`).
+- `complete_task` with `modelAlias`, token totals, outcome.
+
+### Git
+
+- One worktree per brief under `WORKER_WORKTREE_BASE`.
+- Branch naming convention documented for Epic 8 PR linking.
+
+### PR output
+
+- Open draft PR via forge adapter from connected repo (Epic 3).
+
+## Subtasks
+
+- [ ] Wire live dequeue loop (non-dry-run) end-to-end
+- [ ] Role prompt = `.cursor/agents/role-*.md` + brief markdown path
+- [ ] Worktree create/teardown on success and failure
+- [ ] Draft PR via forge adapter when implementer completes
+- [ ] `docs/worker.md` matches shipped behavior
+
+## Estimation
+
+L
+
+## Acceptance criteria
+
+- [ ] One implementer brief from `.convoys/` completes unattended on a connected repo
+- [ ] `complete_task` includes model alias + token totals
+- [ ] Failed runs tear down worktree and mark run `failed`
+- [ ] `WORKER_DRY_RUN=1` still exits cleanly without Redis
+
+## Links
+
+- Epic: `./Epic-apps-worker-v1.md`
+- Plan: `../Plan-automation-foundation.md`
+- Design reference: Autensa convoy build phase (concepts only — no OpenClaw dependency)
--- a/plans/Plan-automation-foundation/Epic-apps-worker-v1/Task-worker-operator-notes.md
+++ b/plans/Plan-automation-foundation/Epic-apps-worker-v1/Task-worker-operator-notes.md
@ -0,0 +1,73 @@
+---
+kind: task
+slug: worker-operator-notes
+title: Operator notes — mid-build context injection
+plan_slug: automation-foundation
+epic_slug: apps-worker-v1
+status: draft
+priority: P2
+tenant_id: global
+owner: unassigned
+cursor_todo_id: null
+updated_at: "2026-08-03"
+---
+
+# Task summary
+
+Let operators send **queued notes** and **direct messages** to an active worker session without waiting for a human gate.
+
+## Description
+
+**Benchmark:** Autensa Operator Chat — two delivery modes:
+
+| Mode | Delivery | Use case |
+|------|----------|----------|
+| **Queued note** | Injected at next checkpoint / tool round | "Use the existing auth middleware" |
+| **Direct message** | Prepended to very next model call | Urgent course correction |
+
+Echodo scope is convoy-scoped (not per arbitrary task board). Gate co-pilot (Epic 5) stays gate-focused; this extends intervention to `in_progress` convoys.
+
+### Data model
+
+`convoy_operator_messages` table (workspace-scoped):
+
+- `convoy_id`, `brief_id` (optional), `agent_run_id` (optional)
+- `delivery`: `queued` | `direct`
+- `body`, `author_user_id`, `delivered_at`, `created_at`
+
+### APIs
+
+- tRPC `convoys.postOperatorNote` (workspace member)
+- Worker polls or receives note via job metadata on each tool round
+
+### UI
+
+- Mission control detail panel: chat thread when convoy status is `in_progress`
+- Distinct from gate co-pilot (separate tab or section)
+
+## Dependencies
+
+- [Task-worker-minimal-tool-loop](./Task-worker-minimal-tool-loop.md)
+- Epic 5 mission control detail panel (extend, do not rewrite)
+
+## Subtasks
+
+- [ ] Schema + tRPC mutations/list
+- [ ] Worker consumes queued notes between tool steps
+- [ ] Worker consumes direct messages before next `generateText`
+- [ ] Mission control UI thread for `in_progress` convoys
+
+## Estimation
+
+M
+
+## Acceptance criteria
+
+- [ ] Operator posts queued note; worker sees it on next tool round
+- [ ] Direct message arrives before next model call when run is active
+- [ ] Messages are workspace-scoped and auditable in convoy detail
+
+## Links
+
+- Epic: `./Epic-apps-worker-v1.md`
+- Related: [Task-mission-control-ui](../Epic-mission-control-gates-ui/Task-mission-control-ui.md)
--- a/plans/Plan-automation-foundation/Epic-connected-repos/Epic-connected-repos.md
+++ b/plans/Plan-automation-foundation/Epic-connected-repos/Epic-connected-repos.md
@ -7,7 +7,7 @@ status: done
 priority: P1
 tenant_id: global
 cursor_epic_id: null
-updated_at: "2026-07-03"
+updated_at: "2026-08-03"
 ---

 # Epic objective
@ -49,3 +49,9 @@ See automation foundation discussion: Vercel/Coolify-style “show me your repo.
 - [x] [Task-db-connected-repos-migration](./Task-db-connected-repos-migration.md)
 - [x] [Task-forge-adapters-trpc](./Task-forge-adapters-trpc.md)
 - [x] [Task-settings-connected-repos-ui](./Task-settings-connected-repos-ui.md)
+
+## Follow-on tasks (Autensa benchmark)
+
+| Task | Link |
+|------|------|
+| Repo setup preflight tab | [Task-repo-setup-preflight](./Task-repo-setup-preflight.md) |
--- a/plans/Plan-automation-foundation/Epic-connected-repos/Task-repo-setup-preflight.md
+++ b/plans/Plan-automation-foundation/Epic-connected-repos/Task-repo-setup-preflight.md
@ -0,0 +1,70 @@
+---
+kind: task
+slug: repo-setup-preflight
+title: Connected repo setup preflight — readiness before automation
+plan_slug: automation-foundation
+epic_slug: connected-repos
+status: draft
+priority: P2
+tenant_id: global
+owner: unassigned
+cursor_todo_id: null
+updated_at: "2026-08-03"
+---
+
+# Task summary
+
+**Repo Setup** tab on connected repos: verify a remote is automation-ready before orchestrator/worker dispatch (git access, default branch, CI permissions, required secrets).
+
+## Description
+
+**Benchmark:** Autensa v2.5.1 Repo Setup tab — preflight checks with in-UI fixes where the forge API allows. Echodo adapts this to multitenant `repo_links` and forge adapters (Forgejo + GitHub minimum).
+
+### Preflight checks
+
+| Check | Pass criteria |
+|-------|---------------|
+| Git clone | Shallow clone with stored credentials |
+| Default branch | API-detected default matches stored value |
+| Forge API | PR list/create permission |
+| CI visibility | Latest workflow run status readable |
+| Actions token (GitHub) | `contents` + `pull-requests` write if applicable |
+| Required secrets | Names present (values never stored in Echodo DB) |
+
+### UI
+
+- Settings → Connected repos → per-repo **Setup** tab
+- Traffic-light checklist with expandable failure reasons
+- Block convoy **Go** when critical checks fail (soft warning vs hard block — configurable per workspace)
+
+### Fixes (where API supports)
+
+- GitHub: suggest/set workflow token permissions; link to docs for secret creation (do not write secret values through Echodo — operator pastes in forge UI)
+
+## Dependencies
+
+- Epic 3 connected repos (shipped)
+- Epic 6 orchestrator should read preflight status before enqueue
+
+## Subtasks
+
+- [ ] `repoPreflightStatus` zod schema in `packages/shared`
+- [ ] Forge adapter methods: `getDefaultBranch`, `listWorkflowRuns`, `getRepoPermissions`
+- [ ] tRPC `connectedRepos.runPreflight`
+- [ ] Settings UI Setup tab
+- [ ] Orchestrator skips dispatch when repo link `preflight_ok` is false
+
+## Estimation
+
+L
+
+## Acceptance criteria
+
+- [ ] Repo with wrong default branch shows failing check with detected branch name
+- [ ] Private repo without token shows auth failure (no silent pass)
+- [ ] Preflight re-runnable without reconnecting repo
+- [ ] No forge secret values persisted in Postgres
+
+## Links
+
+- Epic: `./Epic-connected-repos.md`
--- a/plans/Plan-automation-foundation/Epic-fleet-dashboard/Epic-fleet-dashboard.md
+++ b/plans/Plan-automation-foundation/Epic-fleet-dashboard/Epic-fleet-dashboard.md
@ -7,14 +7,16 @@ status: draft
 priority: P2
 tenant_id: global
 cursor_epic_id: null
-updated_at: "2026-07-03"
+updated_at: "2026-08-03"
 ---

 # Epic objective

 Top-level **fleet** view: all `convoys-<repo>` workspaces (or connected repos) in one screen. Roll up convoys by status, gate inbox count, active `agent_runs`, recent spend by model alias.

-Answers: “What needs me across all projects?”
+Live **activity feed** (SSE) from `convoy_events` — Autensa benchmark for operator situational awareness.
+
+Answers: “What needs me across all projects?” and “What just happened?”

 ## In scope / out of scope

@ -23,10 +25,12 @@ Answers: “What needs me across all projects?”
 - Route `/fleet` or home dashboard for operator account.
 - Cross-workspace tRPC (membership-scoped).
 - Links into per-repo mission control (Epic 5).
+- SSE activity stream (workspace + fleet scope).

 **Out of scope**

 - Multi-tenant fleet across unrelated customers (operator's own repos only for v1).
+- Autopilot research/ideation feeds.

 ## Dependencies

@ -36,7 +40,11 @@ Answers: “What needs me across all projects?”
 ## Acceptance criteria

 - [ ] Operator with two connected repos sees combined “3 awaiting approval” badge.
+- [ ] New convoy event appears in activity feed without page refresh.

 ## Tasks

-_To be decomposed when epic graduates to `ready`._
+| Task | Link | Status |
+|------|------|--------|
+| Fleet rollup dashboard | [Task-fleet-rollup-dashboard](./Task-fleet-rollup-dashboard.md) | ready |
+| Live activity feed (SSE) | [Task-convoy-activity-feed-sse](./Task-convoy-activity-feed-sse.md) | draft |
--- a/plans/Plan-automation-foundation/Epic-fleet-dashboard/Task-convoy-activity-feed-sse.md
+++ b/plans/Plan-automation-foundation/Epic-fleet-dashboard/Task-convoy-activity-feed-sse.md
@ -0,0 +1,63 @@
+---
+kind: task
+slug: convoy-activity-feed-sse
+title: Live convoy activity feed (SSE)
+plan_slug: automation-foundation
+epic_slug: fleet-dashboard
+status: draft
+priority: P2
+tenant_id: global
+owner: unassigned
+cursor_todo_id: null
+updated_at: "2026-08-03"
+---
+
+# Task summary
+
+Real-time activity stream from `convoy_events` for mission control and fleet dashboard — research/build/gate/health events without full page refresh.
+
+## Description
+
+**Benchmark:** Autensa live activity feed (SSE). Echodo already appends `convoy_events` for role completion, transitions, and (future) health events — this task exposes them as a **subscription surface**.
+
+### Transport
+
+- Prefer Next.js Route Handler `GET /api/convoys/activity/stream` with SSE (workspace-scoped via session)
+- Alternative: tRPC subscription if infra already supports it — pick one, document in ADR addendum
+
+### Event types surfaced
+
+- `role_complete`, `transition`, `health`, `gate_action`
+- Payload: `convoy_id`, `title`, `event_type`, `role`, `outcome`, `occurred_at`, `estimated_cost_usd` (when present)
+
+### UI
+
+- Mission control: collapsible **Activity** strip (last 20 events for current workspace)
+- Fleet dashboard: global feed with product/repo filter
+- Fallback: React Query poll every 30s when SSE unavailable
+
+## Dependencies
+
+- Epic 4 `convoy_events` (shipped)
+- [Task-orchestrator-auto-nudge](../Epic-apps-orchestrator/Task-orchestrator-auto-nudge.md) for health events (soft — feed works without it)
+
+## Subtasks
+
+- [ ] SSE route with workspace membership check
+- [ ] `convoy_events` tail query + fan-out on insert (LISTEN/NOTIFY or short poll server-side)
+- [ ] `ActivityFeed` client component
+- [ ] Rate limit / connection cap per user
+
+## Estimation
+
+M
+
+## Acceptance criteria
+
+- [ ] New `role_complete` event appears in mission control within 5s without refresh
+- [ ] User cannot subscribe to another workspace's events
+- [ ] Connection closes cleanly on navigation
+
+## Links
+
+- Epic: `./Epic-fleet-dashboard.md`
--- a/plans/Plan-automation-foundation/Epic-fleet-dashboard/Task-fleet-rollup-dashboard.md
+++ b/plans/Plan-automation-foundation/Epic-fleet-dashboard/Task-fleet-rollup-dashboard.md
@ -0,0 +1,58 @@
+---
+kind: task
+slug: fleet-rollup-dashboard
+title: Fleet dashboard — cross-repo convoy rollup
+plan_slug: automation-foundation
+epic_slug: fleet-dashboard
+status: ready
+priority: P2
+tenant_id: global
+owner: unassigned
+cursor_todo_id: null
+updated_at: "2026-08-03"
+---
+
+# Task summary
+
+Operator home at `/fleet` (or authenticated landing): rollup gate inbox counts, convoys by status, and active `agent_runs` across all workspaces the user belongs to.
+
+## Description
+
+Core Epic 9 scope. Answers: **"What needs me across all projects?"**
+
+### UI
+
+- Card per connected repo / convoy workspace
+- Badge: N awaiting approval (sum of `GATE_INBOX_STATUSES`)
+- Quick links into `/{workspaceSlug}/mission-control?convoy=...`
+- Active runs count + blocked count
+
+### API
+
+- tRPC `fleet.summary` — membership-scoped cross-workspace query (no client-provided workspace_id trust)
+- Reuse `MISSION_CONTROL_KANBAN_COLUMNS` counts per workspace
+
+## Dependencies
+
+- Epic 4, 5 (shipped)
+
+## Subtasks
+
+- [ ] `fleet` tRPC router with membership joins
+- [ ] `/fleet` route (app layout)
+- [ ] Icon rail / top-header link when user has ≥2 workspaces
+- [ ] Empty state for single-workspace operators
+
+## Estimation
+
+M
+
+## Acceptance criteria
+
+- [ ] Operator with two connected repos sees combined "3 awaiting approval" badge
+- [ ] Clicking badge navigates to correct workspace mission control with inbox focused
+- [ ] Query performs one round-trip (no N+1 per workspace in UI)
+
+## Links
+
+- Epic: `./Epic-fleet-dashboard.md`
--- a/plans/Plan-automation-foundation/Epic-forge-webhooks/Epic-forge-webhooks.md
+++ b/plans/Plan-automation-foundation/Epic-forge-webhooks/Epic-forge-webhooks.md
@ -7,13 +7,15 @@ status: draft
 priority: P2
 tenant_id: global
 cursor_epic_id: null
-updated_at: "2026-07-03"
+updated_at: "2026-08-03"
 ---

 # Epic objective

 Webhook receivers on Echodo for connected forges: PR opened → `pr_review`; PR merged → trigger doc-writer job + `merged`; CI failed → optional `blocked` or notify. Supports Forgejo and GitHub adapters minimum.

+PR checks recovery UI in mission control when CI fails (Autensa benchmark).
+
 ## In scope / out of scope

 **In scope**
@ -21,6 +23,7 @@ Webhook receivers on Echodo for connected forges: PR opened → `pr_review`; PR
 - `POST /api/webhooks/forge/:repoLinkId` with signature verification.
 - Map events to `transition_convoy_status(actor: automation)`.
 - Link PR branch to convoy/brief via branch naming convention or PR body marker.
+- PR checks recovery panel (classify failures, rerun where API allows).

 **Out of scope**

@ -33,7 +36,11 @@ Webhook receivers on Echodo for connected forges: PR opened → `pr_review`; PR
 ## Acceptance criteria

 - [ ] Merge on Forgejo moves convoy from `pr_review` to `merged` when Gate 2 approved.
+- [ ] Mission control shows PR check failures with recovery actions for `pr_review` convoys.

 ## Tasks

-_To be decomposed when epic graduates to `ready`._
+| Task | Link | Status |
+|------|------|--------|
+| Webhook receiver | [Task-forge-webhook-receiver](./Task-forge-webhook-receiver.md) | ready |
+| PR checks recovery panel | [Task-pr-checks-recovery-panel](./Task-pr-checks-recovery-panel.md) | draft |
--- a/plans/Plan-automation-foundation/Epic-forge-webhooks/Task-forge-webhook-receiver.md
+++ b/plans/Plan-automation-foundation/Epic-forge-webhooks/Task-forge-webhook-receiver.md
@ -0,0 +1,59 @@
+---
+kind: task
+slug: forge-webhook-receiver
+title: Forge webhook receiver — PR and CI events
+plan_slug: automation-foundation
+epic_slug: forge-webhooks
+status: ready
+priority: P2
+tenant_id: global
+owner: unassigned
+cursor_todo_id: null
+updated_at: "2026-08-03"
+---
+
+# Task summary
+
+`POST /api/webhooks/forge/:repoLinkId` with HMAC signature verification; map forge events to `transition_convoy_status(actor: automation)`.
+
+## Description
+
+Core Epic 8 scope. Link PRs to convoys via branch naming convention (`convoy/<slug>/brief-N`) or PR body marker (`echodo-convoy: <uuid>`).
+
+### Events (minimum)
+
+| Forge event | Convoy transition |
+|-------------|-------------------|
+| PR opened | Ensure convoy in `pr_review` (if Gate 2 not yet human-approved, notify only) |
+| PR merged | `pr_review` → `merged` when Gate 2 already approved |
+| CI completed (failure) | Optional `blocked` or `convoy_events` notify |
+
+### Security
+
+- Per-repo webhook secret in `repo_links` metadata (reference env var name, not value)
+- Zod-validated payload per `forge_type`
+
+## Dependencies
+
+- Epic 3, 4, 7
+
+## Subtasks
+
+- [ ] Route handler + signature verification
+- [ ] GitHub + Forgejo payload parsers
+- [ ] Convoy lookup by branch/body marker
+- [ ] Register webhook URL helper on connect (optional UX)
+
+## Estimation
+
+M
+
+## Acceptance criteria
+
+- [ ] Merge webhook moves convoy from `pr_review` to `merged` when Gate 2 approved
+- [ ] Invalid signature returns 401
+- [ ] Unknown branch logs and returns 200 (no retry storm)
+
+## Links
+
+- Epic: `./Epic-forge-webhooks.md`
--- a/plans/Plan-automation-foundation/Epic-forge-webhooks/Task-pr-checks-recovery-panel.md
+++ b/plans/Plan-automation-foundation/Epic-forge-webhooks/Task-pr-checks-recovery-panel.md
@ -0,0 +1,62 @@
+---
+kind: task
+slug: pr-checks-recovery-panel
+title: PR checks recovery panel in mission control
+plan_slug: automation-foundation
+epic_slug: forge-webhooks
+status: draft
+priority: P2
+tenant_id: global
+owner: unassigned
+cursor_todo_id: null
+updated_at: "2026-08-03"
+---
+
+# Task summary
+
+When a convoy is in `pr_review`, show linked PR CI status in mission control with classified failures and recovery actions (rerun jobs, link to repo setup).
+
+## Description
+
+**Benchmark:** Autensa v2.5.1 PR checks recovery panel — classify failures as retryable, repo setup, or external provider; offer rerun where forge API supports it.
+
+### UI (convoy detail panel)
+
+- PR link + check run list (name, status, conclusion)
+- Failure classification badge per check
+- Actions: **Rerun failed jobs** (GitHub Actions API), **Open repo setup** (Epic 3 follow-on task)
+
+### Backend
+
+- tRPC `convoys.getPrChecks` — workspace-scoped, uses forge adapter
+- `convoys.rerunFailedChecks` — admin only; audit log in `convoy_events`
+
+### Webhook integration
+
+- CI `completed` webhook refreshes cached check status on convoy metadata
+
+## Dependencies
+
+- [Task-forge-webhook-receiver](./Task-forge-webhook-receiver.md)
+- [Task-repo-setup-preflight](../Epic-connected-repos/Task-repo-setup-preflight.md) (link target for setup failures)
+
+## Subtasks
+
+- [ ] Forge adapter: list check runs for PR
+- [ ] Failure classifier (heuristics: permissions → setup, timeout → retryable, external → informational)
+- [ ] Mission control PR checks section when status is `pr_review`
+- [ ] Rerun action + event logging
+
+## Estimation
+
+M
+
+## Acceptance criteria
+
+- [ ] Failed lint CI shows retryable with working rerun (GitHub fixture or manual dogfood)
+- [ ] Missing Actions permission links to repo setup tab with failing check highlighted
+- [ ] No recovery action exposed without workspace admin role
+
+## Links
+
+- Epic: `./Epic-forge-webhooks.md`
--- a/plans/Plan-automation-foundation/Epic-mission-control-gates-ui/Epic-mission-control-gates-ui.md
+++ b/plans/Plan-automation-foundation/Epic-mission-control-gates-ui/Epic-mission-control-gates-ui.md
@ -7,7 +7,7 @@ status: done
 priority: P1
 tenant_id: global
 cursor_epic_id: null
-updated_at: "2026-07-03"
+updated_at: "2026-08-03"
 ---

 # Epic objective
@ -48,3 +48,12 @@ Echodo UI for human gates and convoy visibility:

 - [x] [Task-trpc-convoys-router](./Task-trpc-convoys-router.md)
 - [x] [Task-mission-control-ui](./Task-mission-control-ui.md)
+
+## Follow-on tasks (Autensa benchmark)
+
+| Task | Link |
+|------|------|
+| Convoy DAG visualization | [Task-convoy-dag-visualization](./Task-convoy-dag-visualization.md) |
+| Automation tiers (supervised / semi-auto) | [Task-automation-tiers](./Task-automation-tiers.md) |
+
+Operator mid-build chat is tracked under Epic 7: [Task-worker-operator-notes](../Epic-apps-worker-v1/Task-worker-operator-notes.md).
--- a/plans/Plan-automation-foundation/Epic-mission-control-gates-ui/Task-automation-tiers.md
+++ b/plans/Plan-automation-foundation/Epic-mission-control-gates-ui/Task-automation-tiers.md
@ -0,0 +1,68 @@
+---
+kind: task
+slug: automation-tiers
+title: Automation tiers — supervised, semi-auto, full-auto gate policy
+plan_slug: automation-foundation
+epic_slug: mission-control-gates-ui
+status: draft
+priority: P2
+tenant_id: global
+owner: unassigned
+cursor_todo_id: null
+updated_at: "2026-08-03"
+---
+
+# Task summary
+
+Per-convoy (or per-workspace default) **automation tier** controlling which human gates auto-advance when CI and review agents pass.
+
+## Description
+
+**Benchmark:** Autensa automation tiers (Supervised / Semi-Auto / Full Auto). Maps onto Echodo's explicit gate model without removing human override.
+
+| Tier | Gate 1 (plan) | Gate 2 (merge) | Gate 3 (prod) |
+|------|---------------|----------------|---------------|
+| **Supervised** | Human | Human | Human |
+| **Semi-auto** | Human | Auto when CI green + review agent pass | Human |
+| **Full-auto** | Human | Auto when CI green | Auto when deploy webhook fires (future) |
+
+### Schema
+
+- `convoys` metadata field `automation_tier`: `supervised` | `semi_auto` | `full_auto` (default `supervised`)
+- Workspace default in settings (optional)
+
+### Behavior
+
+- Orchestrator / webhook handlers consult tier before calling `transition_convoy_status` with `actor: automation`
+- Mission control shows tier badge on convoy card; editable before **Go**
+
+### Out of scope
+
+- Auto-approving Gate 1 (plan always human in v1)
+- Autopilot research/ideation (Autensa-specific)
+
+## Dependencies
+
+- Epic 5 gates UI
+- Epic 8 webhooks for semi-auto Gate 2
+
+## Subtasks
+
+- [ ] Zod enum in `packages/shared/convoy-lifecycle`
+- [ ] Create convoy dialog + detail panel tier selector
+- [ ] Orchestrator respects tier on CI webhook path
+- [ ] Document tier semantics in `docs/orchestrator.md`
+
+## Estimation
+
+S
+
+## Acceptance criteria
+
+- [ ] Semi-auto convoy auto-transitions `pr_review` → `merged` on merge webhook only when CI was green at merge time
+- [ ] Supervised convoy never auto-merges
+- [ ] Tier visible on kanban card tooltip
+
+## Links
+
+- Epic: `./Epic-mission-control-gates-ui.md`
--- a/plans/Plan-automation-foundation/Epic-mission-control-gates-ui/Task-convoy-dag-visualization.md
+++ b/plans/Plan-automation-foundation/Epic-mission-control-gates-ui/Task-convoy-dag-visualization.md
@ -0,0 +1,58 @@
+---
+kind: task
+slug: convoy-dag-visualization
+title: Convoy DAG visualization for parallel briefs
+plan_slug: automation-foundation
+epic_slug: mission-control-gates-ui
+status: draft
+priority: P2
+tenant_id: global
+owner: unassigned
+cursor_todo_id: null
+updated_at: "2026-08-03"
+---
+
+# Task summary
+
+Read-only dependency graph in mission control showing brief/subtask relationships and `multitask_group` parallel lanes.
+
+## Description
+
+**Benchmark:** Autensa Convoy Mode DAG — visual dependency graph for parallel subtasks. Echodo already has `multitask_group` on `convoy_events` and parallel implementer fan-out in the orchestrator ADR; this task adds **visibility**, not new scheduling logic.
+
+### Data source
+
+- Parse brief frontmatter / `slice_dependencies` from `.convoys/<slug>/brief-*.md` (via convoy object metadata or file reconcile)
+- Overlay live status from `agent_runs` + convoy status per brief
+
+### UI
+
+- Mission control detail panel tab: **Graph**
+- Nodes = briefs; edges = dependencies; color = run status (pending, running, done, blocked)
+- Click node selects brief in detail panel
+
+## Dependencies
+
+- Epic 5 mission control (shipped)
+- Epic 6 orchestrator parallel dispatch (shipped)
+
+## Subtasks
+
+- [ ] Shared zod type `ConvoyDag` in `packages/shared`
+- [ ] tRPC `convoys.getDag`
+- [ ] React graph component (lightweight — e.g. layered layout, no heavy graph lib unless justified)
+- [ ] Empty state when single-brief convoy
+
+## Estimation
+
+M
+
+## Acceptance criteria
+
+- [ ] Two-brief convoy with dependency shows correct edge direction
+- [ ] Parallel briefs without dependency appear in same column/lane
+- [ ] Blocked brief visually distinct from pending
+
+## Links
+
+- Epic: `./Epic-mission-control-gates-ui.md`
--- a/plans/Plan-automation-foundation/Epic-model-provider-registry/Epic-model-provider-registry.md
+++ b/plans/Plan-automation-foundation/Epic-model-provider-registry/Epic-model-provider-registry.md
@ -7,7 +7,7 @@ status: done
 priority: P1
 tenant_id: global
 cursor_epic_id: null
-updated_at: "2026-07-03"
+updated_at: "2026-08-03"
 agent_prompt: |
  Epic: model provider registry (docs/adr-model-routing.md). Server-only change —
  no UI component library inventing; reuse components/ui/. Admin-only mutations.
@ -32,12 +32,9 @@ This epic is the first shipped piece of the automation-foundation plan and unblo
 - `complete_task` optional model logging fields.
 - `.env.example` env var documentation.

-**Out of scope**
+**Out of scope (moved to follow-on task)**

- Orchestrator dispatch (consumes resolver later).
- Worker inference loop.
- Parsing `model_policy` from `.convoys/*.md` files (orchestrator reads later).
- Billing/budget enforcement.
+- Billing/budget enforcement → [Task-cost-budget-ui-and-caps](./Task-cost-budget-ui-and-caps.md)

 ## Related tasks

@ -47,9 +44,15 @@ This epic is the first shipped piece of the automation-foundation plan and unblo
 | DB schema and migration | `./Task-db-model-providers-migration.md` |
 | tRPC models router | `./Task-trpc-models-router.md` |
 | Settings — providers UI | `./Task-settings-ai-providers-ui.md` |
-| Settings — role routing UI | `./Task-settings-role-routing-ui.md` |
+| Settings — role routing UI | `./Task-settings-ai-role-routing-ui.md` |
 | agent_runs model fields + MCP | `./Task-agent-runs-model-logging.md` |

+## Follow-on tasks (Autensa benchmark)
+
+| Task | Link |
+|------|------|
+| Cost visibility & dispatch budget caps | [Task-cost-budget-ui-and-caps](./Task-cost-budget-ui-and-caps.md) |
+
 ## Dependencies

 - Depends on: existing auth + workspace admin patterns in `apps/web`.
@ -57,11 +60,11 @@ This epic is the first shipped piece of the automation-foundation plan and unblo

 ## Acceptance criteria

- [ ] Admin can create `claude-planning` (anthropic) and `qwen-coder` (openai_compatible + base URL) in a workspace.
- [ ] Admin can map `role-implementer` → `qwen-coder`, `role-architect` → `claude-planning`.
- [ ] `testProvider` returns success/failure without exposing API keys.
- [ ] `resolveModel()` unit tests cover override order and missing-alias error.
- [ ] At least one task in this epic completed via `claim_task` → `complete_task` with model metadata.
+- [x] Admin can create `claude-planning` (anthropic) and `qwen-coder` (openai_compatible + base URL) in a workspace.
+- [x] Admin can map `role-implementer` → `qwen-coder`, `role-architect` → `claude-planning`.
+- [x] `testProvider` returns success/failure without exposing API keys.
+- [x] `resolveModel()` unit tests cover override order and missing-alias error.
+- [x] At least one task in this epic completed via `claim_task` → `complete_task` with model metadata.

 ## Proposed timeline

@ -71,6 +74,7 @@ This epic is the first shipped piece of the automation-foundation plan and unblo
 | API | Session 3 | tRPC router |
 | UI | Session 4–5 | Both settings pages |
 | Logging | Session 6 | agent_runs + MCP |
+| Usage & caps | Future | Follow-on task |

 ## Pipeline / convoy mapping

--- a/plans/Plan-automation-foundation/Epic-model-provider-registry/Task-cost-budget-ui-and-caps.md
+++ b/plans/Plan-automation-foundation/Epic-model-provider-registry/Task-cost-budget-ui-and-caps.md
@ -0,0 +1,63 @@
+---
+kind: task
+slug: cost-budget-ui-and-caps
+title: Cost visibility and dispatch budget caps
+plan_slug: automation-foundation
+epic_slug: model-provider-registry
+status: draft
+priority: P2
+tenant_id: global
+owner: unassigned
+cursor_todo_id: null
+updated_at: "2026-08-03"
+---
+
+# Task summary
+
+Surface spend from `convoy_events.estimated_cost_usd` and `agent_runs` token metadata; add optional workspace budget caps that pause orchestrator dispatch when exceeded.
+
+## Description
+
+**Benchmark:** Autensa per-task/product cost tracking and daily/monthly budget caps. Epic 1 deferred billing — this follow-on uses data already on `convoy_events` and run completion.
+
+### Read path (UI)
+
+- Settings → AI → **Usage & budgets**
+- Rollups: today / this month by `model_alias`, by convoy, by role
+- Mission control convoy detail: spend to date for selected convoy
+
+### Write path (enforcement)
+
+- `workspace_budget_caps` table: `period` (`daily` | `monthly`), `cap_usd`, `action` (`pause_dispatch` | `warn_only`)
+- Orchestrator tick: sum `estimated_cost_usd` for period; if over cap and `pause_dispatch`, skip new `role_dispatch` enqueue and write health event
+
+### Cost estimation
+
+- Worker/orchestrator populate `estimated_cost_usd` on `convoy_events.role_complete` using provider-reported usage × static price table in shared config (v1 — no live pricing API)
+
+## Dependencies
+
+- Epic 1 model registry + `convoy_events` cost column (shipped)
+- Epic 7 worker records token totals on complete
+
+## Subtasks
+
+- [ ] Price table stub in `packages/shared` (document as operator-editable later)
+- [ ] tRPC usage queries (workspace-scoped aggregates)
+- [ ] Settings usage page
+- [ ] Budget cap schema + orchestrator guard
+
+## Estimation
+
+M
+
+## Acceptance criteria
+
+- [ ] Settings shows month-to-date spend grouped by model alias
+- [ ] With cap set and exceeded, orchestrator logs skip reason and does not enqueue new jobs
+- [ ] Warn-only cap emits event but does not block dispatch
+
+## Links
+
+- Epic: `./Epic-model-provider-registry.md`
+- Unblocks: Epic 9 fleet spend rollup
--- a/plans/Plan-automation-foundation/Plan-automation-foundation.md
+++ b/plans/Plan-automation-foundation/Plan-automation-foundation.md
@ -6,7 +6,7 @@ status: ready
 priority: P1
 tenant_id: global
 cursor_plan_id: null
-updated_at: "2026-07-07"
+updated_at: "2026-08-03"
 agent_prompt: |
  You are implementing Echodo's automation foundation. Read docs/adr-model-routing.md
  before touching model routing code. This repo uses pnpm, Drizzle, tRPC, zod in
@ -28,7 +28,7 @@ agent_prompt: |

 Echodo is evolving from a multitenant task app with MCP coordination into **mission control** for an agent pipeline: fleet view, human gates, per-role model routing, and (later) a built-in orchestrator and worker. This plan delivers the **first concrete slice** — the model provider registry and role routing system defined in [docs/adr-model-routing.md](../../docs/adr-model-routing.md).

-Later epics in this plan (2–9) add convoy scaffolding, connected repos, lifecycle MCP, gates UI, orchestrator, worker, webhooks, and fleet dashboard. Epic 1 gives them a stable `resolveModel()` contract.
+Epics 1–6 and the mission-control UI are **shipped**. Epic 7 (worker) is **in progress** — skeleton exists in `apps/worker`. Epics 8–9 remain draft with decomposed tasks. Follow-on tasks from the [Autensa benchmark](#autensa-benchmark-2026-08) are injected into epics 1, 3, 5, 6, 7, 8, and 9 (concepts only — no OpenClaw dependency).

 ## Dogfooding the agent pipeline

@ -115,7 +115,8 @@ Start with **`Task-shared-schemas-resolve-model`** — no DB migration, pure `pa
 **In scope (full plan)**

 - Nine epics — see [Epic roadmap](#epic-roadmap-complete) below.
- Epic 1 (model registry) is **ready** with six tasks; epics 2–9 are **draft** stubs for architect decomposition.
+- Epics 1–6 **done**; Epic 7 **in progress**; Epics 8–9 **draft** with bead-scale tasks.
+- [Autensa benchmark](#autensa-benchmark-2026-08) follow-ons: reliability (checkpoints, auto-nudge), operator UX (activity feed, DAG, PR recovery), cost caps.
 - ADR at `docs/adr-model-routing.md`.

 **Out of scope (separate future plans)**
@ -133,7 +134,7 @@ Start with **`Task-shared-schemas-resolve-model`** — no DB migration, pure `pa

 ## Epic roadmap (complete)

-Sequential order with parallel opportunities noted. Only **Epic 1** has bead-scale tasks today; epics 2–9 are draft stubs ready for architect decomposition when the prior epic lands.
+Sequential order with parallel opportunities noted. **Current focus: Epic 7** (worker end-to-end dogfood), then Epic 8 → 9. Follow-on tasks on done epics can run in parallel when they do not block the critical path.

 ```mermaid
 flowchart LR
@ -162,15 +163,15 @@ flowchart LR

 | # | Epic | Status | Delivers | Depends on |
 |---|------|--------|----------|------------|
-| 1 | [Model provider registry](./Epic-model-provider-registry/Epic-model-provider-registry.md) | **ready** (6 tasks) | `resolveModel()`, Settings AI pages, run model logging | — |
+| 1 | [Model provider registry](./Epic-model-provider-registry/Epic-model-provider-registry.md) | done (+1 follow-on) | `resolveModel()`, Settings AI pages, run model logging; cost caps task | — |
 | 2 | [Pipeline convoy scaffolding](./Epic-pipeline-convoy-scaffolding/Epic-pipeline-convoy-scaffolding.md) | done | `.convoys/`, `log-convoy-event.sh`, `convoy-planning.mdc` — dogfood metrics | — (parallel with 1) |
-| 3 | [Connected repos](./Epic-connected-repos/Epic-connected-repos.md) | done | `repo_links` table, “paste remote URL”, forge adapter interface | — |
+| 3 | [Connected repos](./Epic-connected-repos/Epic-connected-repos.md) | done (+1 follow-on) | `repo_links` table, paste URL, forge adapters; repo setup preflight | — |
 | 4 | [Convoy lifecycle MCP](./Epic-convoy-lifecycle-mcp/Epic-convoy-lifecycle-mcp.md) | done | `create_convoy`, `transition_convoy_status`, state machine §7, `convoy_events` | 2, 3 |
-| 5 | [Mission control gates UI](./Epic-mission-control-gates-ui/Epic-mission-control-gates-ui.md) | done | Convoy kanban, gate inbox, approve/deny/revise, gate co-pilot chat | 4 |
-| 6 | [Orchestrator](./Epic-apps-orchestrator/Epic-apps-orchestrator.md) | done | `apps/orchestrator`, Redis queue, role dispatch, pause at gates | 1, 3, 4 |
-| 7 | [Worker v1](./Epic-apps-worker-v1/Epic-apps-worker-v1.md) | draft | `apps/worker` minimal tool loop, worktree-per-brief, MCP claim/complete | 1, 3, 6 |
-| 8 | [Forge webhooks](./Epic-forge-webhooks/Epic-forge-webhooks.md) | draft | PR opened/merged/CI webhooks → convoy transitions | 3, 4, 7 |
-| 9 | [Fleet dashboard](./Epic-fleet-dashboard/Epic-fleet-dashboard.md) | draft | Cross-repo home, `convoys-*` workspaces, “needs you” rollup | 4, 5 |
+| 5 | [Mission control gates UI](./Epic-mission-control-gates-ui/Epic-mission-control-gates-ui.md) | done (+2 follow-ons) | Kanban, gate inbox, co-pilot; DAG viz + automation tiers | 4 |
+| 6 | [Orchestrator](./Epic-apps-orchestrator/Epic-apps-orchestrator.md) | done (+1 follow-on) | `apps/orchestrator`, Redis queue, role dispatch; auto-nudge | 1, 3, 4 |
+| 7 | [Worker v1](./Epic-apps-worker-v1/Epic-apps-worker-v1.md) | **in progress** (3 tasks) | Tool loop, checkpoints, operator notes | 1, 3, 6 |
+| 8 | [Forge webhooks](./Epic-forge-webhooks/Epic-forge-webhooks.md) | draft (2 tasks) | Webhooks + PR checks recovery panel | 3, 4, 7 |
+| 9 | [Fleet dashboard](./Epic-fleet-dashboard/Epic-fleet-dashboard.md) | draft (2 tasks) | Cross-repo rollup + SSE activity feed | 4, 5 |

 ### Overlap with existing plans

@ -183,7 +184,26 @@ flowchart LR

 ### End-state (what “done” looks like for the full plan)

-Operator opens Echodo fleet view → starts convoy chat → hits **Go** → orchestrator runs pipeline roles with per-role models → worker executes briefs against connected repos → gates surface in inbox with AI co-pilot → approve → repeat until `released`.
+Operator opens Echodo fleet view → starts convoy chat → hits **Go** → orchestrator runs pipeline roles with per-role models → worker executes briefs against connected repos → gates surface in inbox with AI co-pilot → approve → repeat until `released`. Live activity feed, spend rollups, and PR recovery are visible without leaving mission control.
+
+### Autensa benchmark (2026-08)
+
+Studied [crshdn/mission-control](https://github.com/crshdn/mission-control) (Autensa) as a UX/reliability benchmark — **not** as a runtime dependency. Ideas mapped to tasks:
+
+| Autensa concept | Echodo task | Epic |
+|-----------------|-------------|------|
+| Checkpoint & crash recovery | [Task-worker-checkpoint-resume](./Epic-apps-worker-v1/Task-worker-checkpoint-resume.md) | 7 |
+| Operator chat mid-build | [Task-worker-operator-notes](./Epic-apps-worker-v1/Task-worker-operator-notes.md) | 7 |
+| Agent health / auto-nudge | [Task-orchestrator-auto-nudge](./Epic-apps-orchestrator/Task-orchestrator-auto-nudge.md) | 6 |
+| Repo setup preflight | [Task-repo-setup-preflight](./Epic-connected-repos/Task-repo-setup-preflight.md) | 3 |
+| PR checks recovery | [Task-pr-checks-recovery-panel](./Epic-forge-webhooks/Task-pr-checks-recovery-panel.md) | 8 |
+| Convoy DAG visualization | [Task-convoy-dag-visualization](./Epic-mission-control-gates-ui/Task-convoy-dag-visualization.md) | 5 |
+| Automation tiers | [Task-automation-tiers](./Epic-mission-control-gates-ui/Task-automation-tiers.md) | 5 |
+| Cost tracking & budget caps | [Task-cost-budget-ui-and-caps](./Epic-model-provider-registry/Task-cost-budget-ui-and-caps.md) | 1 |
+| Live activity feed (SSE) | [Task-convoy-activity-feed-sse](./Epic-fleet-dashboard/Task-convoy-activity-feed-sse.md) | 9 |
+| Fleet rollup | [Task-fleet-rollup-dashboard](./Epic-fleet-dashboard/Task-fleet-rollup-dashboard.md) | 9 |
+
+**Explicitly not adopted:** OpenClaw Gateway, autopilot research/ideation/swipe, preference learning, agent skills learner, SQLite single-tenant model.

 ## Epics (links)

@ -212,6 +232,20 @@ Operator opens Echodo fleet view → starts convoy chat → hits **Go** → orch
 | M6 Automation | 6–7 | Orchestrator dispatches one implementer brief unattended |
 | M7 PR loop | 8 | Merge webhook moves convoy to `merged` |
 | M8 Fleet | 9 | Single view across ≥2 connected repos |
+| M9 Reliability | 7 | Worker completes brief; checkpoint resume after kill |
+| M10 Observability | 9 | Activity feed shows role_complete without refresh |
+| M11 Spend control | 1 follow-on | Month-to-date spend + dispatch pause on cap |
+
+## Recommended execution order (post-benchmark)
+
+1. **Epic 7 critical path** — [Task-worker-minimal-tool-loop](./Epic-apps-worker-v1/Task-worker-minimal-tool-loop.md) → live dogfood with orchestrator
+2. **Epic 7 reliability** — checkpoint/resume, then operator notes
+3. **Epic 8** — webhook receiver → PR checks panel (needs worker PR output)
+4. **Epic 6 follow-on** — auto-nudge (pairs with checkpoints)
+5. **Epic 3 follow-on** — repo preflight (before unattended prod dogfood)
+6. **Epic 9** — fleet rollup → SSE feed
+7. **Epic 5 follow-ons** — DAG viz, automation tiers (after Epic 8 for semi-auto)
+8. **Epic 1 follow-on** — cost caps (after worker records token totals)

 ## Risks and mitigations