Files
roboco/docs/workflows/PM.md
T
Renn F 5315e9c72d feat: workflow enforcement, RAG upgrade, and permission fixes
Task Management:
  - Add cancellation safeguards: require valid reason category (duplicate,
    obsolete, blocked_permanently, reassigned, scope_change, stakeholder_request)
  - Protect active work from arbitrary cancellation - must pause/block first
  - Auto-notify PM when task is blocked with ACTION REQUIRED message
  - PM task scan now shows blocked tasks needing their attention

  Permissions:
  - Add VIEW_STATS to Developer, QA, Documenter, Head Marketing KB permissions
  - Aligns code with docs/workflows/PERMISSIONS.md specification

  RAG/Embeddings:
  - Upgrade embedding model from all-MiniLM-L6-v2 to nomic-embed-text-v1.5
  - 768 dimensions with 8K token context (vs 512 tokens)
  - Add per-index chunk sizes: docs=1536, journals=1024, others=512
  - Switch to fixed chunking (semantic chunking loads separate MiniLM model)
  - Add einops dependency required by nomic model
2025-12-29 00:30:18 +01:00

253 lines
9.7 KiB
Markdown

# PM Workflow
## Main PM
```
┌─────────────────────────────────────────────────────────────────────────┐
│ MAIN PM WORKFLOW │
└─────────────────────────────────────────────────────────────────────────┘
1. RECEIVE WORK (from Board/CEO)
2. TRIAGE
├── roboco_task_get(task_id) → Read requirements
├── roboco_task_claim(task_id) → Take ownership
└── roboco_task_start(task_id) → Begin triage work
3. PLAN & BREAKDOWN
│ roboco_task_plan(task_id, approach, steps)
│ roboco_task_progress(task_id, "Planning complete", 20)
│ # REQUIRED: Document your planning decisions
│ roboco_journal_decision({
│ title: "Task breakdown for [feature]",
│ context: "Requirements from board",
│ options: ["Option A", "Option B"],
│ chosen: "Option A",
│ rationale: "Because..."
│ })
4. CREATE SUBTASKS (for Cell PMs)
│ For EACH cell subtask:
│ ┌─────────────────────────────────────────────────────────────────┐
│ │ roboco_task_create({ │
│ │ title: "...", │
│ │ description: "...", │
│ │ team: "backend" | "frontend" | "ux_ui", │
│ │ parent_task_id: main_task_id, │
│ │ status: "backlog", ← STARTS IN BACKLOG │
│ │ assigned_to: "be-pm" ← ASSIGN TO CELL PM │
│ │ }) │
│ └─────────────────────────────────────────────────────────────────┘
5. CREATE SESSION (groups related subtasks)
│ roboco_session_create_for_tasks({
│ title: "Feature X Implementation",
│ task_ids: [subtask_1_id, subtask_2_id, ...]
│ })
6. ACTIVATE SUBTASKS
│ For EACH subtask:
│ ┌─────────────────────────────────────────────────────────────────┐
│ │ roboco_task_activate(subtask_id) │
│ │ │
│ │ STATUS: backlog → pending │
│ │ Now visible to Cell PM in roboco_task_scan() │
│ └─────────────────────────────────────────────────────────────────┘
7. NOTIFY CELL PMs
│ roboco_notify_send({
│ recipient: "be-pm",
│ type: "task_assignment",
│ task_id: subtask_id,
│ message: "New task assigned to your cell"
│ })
8. MONITOR & COORDINATE
│ Loop:
│ ├── roboco_task_scan() → Check subtask statuses
│ ├── roboco_channel_history("pm-all") → Cross-cell coordination
│ ├── roboco_journal_read_team("be-pm") → Read cell PM progress
│ ├── Handle escalations from Cell PMs
│ ├── roboco_task_progress(main_task_id, "X% complete", %)
│ └── roboco_journal_entry({type: "coordination", ...})
9. REFLECT & COMPLETE (when all subtasks done)
│ # REQUIRED: Reflect before completing
│ roboco_journal_reflect({
│ task_id: main_task_id,
│ what_done: "Coordinated X cells, Y subtasks",
│ what_learned: "Cross-cell coordination patterns",
│ what_struggled: "Dependency management"
│ })
│ roboco_task_complete(main_task_id)
DONE
```
## Cell PM (be-pm, fe-pm, ux-pm)
```
┌─────────────────────────────────────────────────────────────────────────┐
│ CELL PM WORKFLOW │
└─────────────────────────────────────────────────────────────────────────┘
1. CHECK NOTIFICATIONS
│ roboco_notify_list()
│ You'll receive automatic notifications when:
│ ├── Documenter completes docs (task auto-assigned to you)
│ ├── Agent submits for PM review (task auto-assigned to you)
│ ├── QA/Documenter substitutes with "task_complete"
│ └── Escalations from your cell
│ roboco_notify_ack(notification_id) # Acknowledge each
2. SCAN FOR WORK
│ roboco_task_scan(team="backend")
│ Look for:
│ ├── Tasks in "pending" assigned to me
│ ├── Tasks in "awaiting_pm_review" (need my approval)
│ └── Any remaining escalations
3. CLAIM TASK
│ roboco_task_claim(task_id)
│ STATUS: pending → claimed
│ ASSIGNED_TO: confirmed as me
4. START & PLAN
│ roboco_task_start(task_id)
│ STATUS: claimed → in_progress
│ roboco_task_plan(task_id, approach, steps)
5. CREATE DEV SUBTASKS
│ For EACH dev subtask:
│ ┌─────────────────────────────────────────────────────────────────┐
│ │ roboco_task_create({ │
│ │ title: "...", │
│ │ description: "...", │
│ │ team: "backend", │
│ │ parent_task_id: my_task_id, │
│ │ status: "backlog", ← STARTS IN BACKLOG │
│ │ assigned_to: "be-dev-1" ← OPTIONAL: pre-assign │
│ │ }) │
│ └─────────────────────────────────────────────────────────────────┘
6. ACTIVATE SUBTASKS
│ roboco_task_activate(subtask_id)
│ STATUS: backlog → pending
│ Subtask inherits parent's session automatically
7. NOTIFY DEVELOPERS
│ roboco_notify_send({
│ recipient: "be-dev-1",
│ type: "task_assignment",
│ task_id: subtask_id,
│ message: "Task ready for you"
│ })
8. MONITOR CELL WORK
│ Loop:
│ ├── roboco_notify_list() # Check for auto-assigned tasks
│ ├── roboco_task_scan(team="backend")
│ ├── Watch for "awaiting_pm_review" tasks
│ ├── Handle blockers/escalations
│ └── roboco_task_progress(my_task_id, "X% complete", %)
9. COMPLETE SUBTASKS (after QA + Docs)
│ When subtask reaches "awaiting_pm_review":
│ ├── Task is auto-assigned to you with notification
│ ├── Review the work
│ └── roboco_task_complete(subtask_id)
10. COMPLETE MY TASK (when all subtasks done)
│ roboco_task_complete(my_task_id)
DONE → Main PM notified
```
## Task Status Transitions (PM perspective)
```
PM CREATES:
backlog ──activate──► pending
DEVELOPER CLAIMS:
pending ──claim──► claimed
DEVELOPER WORKS:
claimed ──start──► in_progress
AFTER QA + DOCS:
awaiting_pm_review ──PM completes──► completed
```
## Using Knowledge Base
PMs have full KB access including indexing:
```python
roboco_kb_search("similar past tasks") # Find related work
roboco_rag_query("how did we solve X?") # AI-generated answers
roboco_journal_read_team("be-dev-1") # Read team journals
# Indexing (PM only)
roboco_kb_index_code(["src/**/*.py"]) # Index code for search
roboco_kb_index_docs(["docs/**/*.md"]) # Index documentation
```
See [KNOWLEDGE_BASE.md](./KNOWLEDGE_BASE.md) for full documentation.
## Key Rules
1. **Tasks start in BACKLOG** - PM setup phase
2. **ACTIVATE before anyone can claim** - backlog → pending
3. **Sessions group related tasks** - create before activating
4. **Subtasks inherit parent session** - no need to create new session
5. **NOTIFY after activation** - roboco_notify_send() REQUIRED
6. **JOURNAL decisions** - roboco_journal_decision() for task breakdowns
7. **READ team journals** - roboco_journal_read_team() for monitoring
8. **REFLECT before complete** - roboco_journal_reflect() REQUIRED
9. **Only PM can COMPLETE** - after full workflow (dev → QA → docs → PM review)