# Task Pool Workflow — Investor Flow ## How Tasks Flow ``` new → research → implement → code_review → complete ``` ## Canonical design docs - `docs/FUNCTIONAL_DESIGN.md` — product status - `docs/TECH_DESIGN.md` — technical status - `CONTEXT.md` — glossary only ## Priority gap tasks (2026-07-18 audit) Pick in this order when available: 1. **P0** `fix-macro-commentary-tests` — restore green suite 2. **P0** `wire-sizing-risk-trpc-m20` — size + risk on tRPC + M20 UI 3. **P1** `frontend-api-client-coverage` — unblock UI tasks 4. **P1** `persist-trade-plan-execution-loop` — multi-tenant process loop 5. **P1** `wire-emotion-logger-backend` 6. **P2** strategy lab, screener UI, thesis/derisk UI, options convexity UI 7. **P3** reports UI, schema strategies dedupe ## For Agents ```bash cd /Users/laptran/Documents/investor-flow python3 ~/.automaton/scripts/status.py --available --agent --project . python3 ~/.automaton/scripts/status.py --claim --task --agent --project . # implement from SPEC.md python3 ~/.automaton/scripts/status.py --transition code_review --task --project . ``` ### Rules - ADR-0007: no unframed imperative trade verbs - Reviewer ≠ implementer - Backend tests: `cd app/server && npm test` - Primary-rule lint: `cd app && node --test --experimental-strip-types src/__tests__/primary-rule-lint.test.ts` ## Historical note Many modules under `.automaton/tasks/complete/` are finished vertical slices. The old "21 task pool" table is obsolete; use the priority list above.