1---2name: pm-router3description: Routes PM coordinator to appropriate skills and sub-agents based on workflow phase, task category, and signal keywords. Use when starting PM session, assigning tasks, running retrospectives, or managing PRD.4---5
6# PM Skill Router
7
8> "Right skill for the right phase at the right time."
9
10---
11
12## Route by Current Status (Post-Completion Flow)
13
14⚠️ **CRITICAL: When in post-completion workflow, follow this routing table to continue through ALL phases!**
15
16| Current Status | Action | Next Skill/Phase |
17| --------------------------------- | ------------------------------------------------------------ | ------------------------------------ |
18| `retrospective_synthesized` | Check if playtest required → Use playtest session skill OR skip to PRD reorganization | `pm-retrospective-playtest-session` or `pm-organization-prd-reorganization` |
19| `playtest_complete` | Continue to PRD reorganization | `pm-organization-prd-reorganization` |
20| `playtest_skipped` | Continue to PRD reorganization | `pm-organization-prd-reorganization` |
21| `prd_refinement` | Continue to cleanup phase | Cleanup (inline in pm-workflow) |
22| `cleanup_completed` | Continue to skill research | `pm-improvement-skill-research` |
23| `skill_updates_applied` | Continue to task selection | `pm-organization-task-selection` |
24| `task_ready` (post-completion) | Create test plan | `pm-planning-test-planning` |
25| `test_plan_ready` | Assign task, send message, EXIT | (Exit - worker has task) |
26
27**Playtest Decision (at `retrospective_synthesized`):**
28
29Skip playtest for:
30- Test infrastructure bugfixes
31- Non-gameplay tasks (CI/CD, tooling, documentation)
32- Backend-only changes without visual impact
33- Small typo/minor text fixes
34
35Use playtest for:
36- Gameplay mechanics (movement, shooting, physics)
37- Visual features (shaders, materials, effects)
38- UI/UX changes (HUD, menus, interactions)
39- Character/weapon behavior changes
40- Multiplayer features
41
42---
43
44## Route by Workflow Phase
45
46| Phase | Primary Skills | Sub-Agents |
47| ---------------------- | -------------------------------------------------------------------- | ------------------------------ |
48| **Startup** | `pm-organization-scale-adaptive` | - |
49| **Task Selection** | `pm-organization-task-selection`, `pm-organization-task-research` | `pm-task-researcher` |
50| **Test Planning** | `pm-planning-test-planning` | `pm-test-planner` |
51| **Assignment** | (use selected skills) | - |
52| **Retrospective** | `pm-retrospective-facilitation`, `pm-retrospective-playtest-session` | `pm-retrospective-facilitator` |
53| **PRD Update** | `pm-organization-prd-reorganization` | `pm-prd-organizer` |
54| **Skill Research** | `pm-improvement-skill-research`, `pm-improvement-self-improvement` | `skill-researcher` |
55| **Architecture Check** | `pm-validation-architecture` | `pm-architecture-validator` |
56
57## Route by Task Category
58
59| Category | Agent | PM Skills to Load |
60| -------------------- | ---------- | --------------------------------------------------------------------- |
61| `architectural` | developer | `pm-organization-task-research`, `pm-validation-architecture` |
62| `integration` | developer | `pm-organization-task-research` |
63| `functional` | developer | `pm-organization-task-research` |
64| `visual` | techartist | `pm-configuration-asset-coordination`, `pm-configuration-vite-assets` |
65| `shader` | techartist | `pm-configuration-asset-coordination` |
66| `polish` | techartist | `pm-configuration-asset-coordination` |
67| Any with multiplayer | developer | `pm-validation-architecture` |
68
69## Route by Signal Keywords
70
71| Signal in Task | Route To |
72| ---------------------------------------------------- | ------------------------------------- |
73| "asset", "model", "fbx", "texture", "gltf" | `pm-configuration-asset-coordination` |
74| "vite", "public/", "src/assets/" | `pm-configuration-vite-assets` |
75| "multiplayer", "server", "colyseus", "authoritative" | `pm-validation-architecture` |
76| "test", "e2e", "validation", "acceptance" | `pm-planning-test-planning` |
77| "gdd", "design", "mechanic" | `pm-organization-prd-reorganization` |
78
79## All PM Skills
80
81| Skill | Purpose | Phase |
82| ------------------------------------- | ------------------------------------- | -------------- |
83| `pm-organization-scale-adaptive` | Adjust planning depth by PRD size | Startup |
84| `pm-organization-task-selection` | Priority algorithm for task selection | Task Selection |
85| `pm-organization-task-research` | Codebase research before assignment | Task Selection |
86| `pm-planning-test-planning` | Collaborative test planning | Test Planning |
87| `pm-configuration-asset-coordination` | Asset coordination for parallel work | Assignment |
88| `pm-configuration-vite-assets` | Vite 6 asset patterns | Assignment |
89| `pm-validation-architecture` | Server-authoritative validation | Assignment |
90| `pm-retrospective-facilitation` | Worker retrospective orchestration | Retrospective |
91| `pm-retrospective-playtest-session` | Game Designer playtest coordination | Retrospective |
92| `pm-organization-prd-reorganization` | GDD-to-PRD task extraction | PRD Update |
93| `pm-improvement-skill-research` | Multi-agent skill improvements | Skill Research |
94| `pm-improvement-self-improvement` | PM self-improvement | Skill Research |
95
96## Sub-Agents (via Task tool)
97
98| Sub-Agent | Model | Purpose |
99| ------------------------------ | ------- | ---------------------------------------- |
100| `pm-task-researcher` | Haiku | Codebase research before task assignment |
101| `pm-test-planner` | Inherit | Test planning with QA+GD |
102| `pm-retrospective-facilitator` | Inherit | Retrospective orchestration |
103| `pm-prd-organizer` | Inherit | PRD reorganization |
104| `pm-architecture-validator` | Haiku | Architecture gap detection |
105| `skill-researcher` | Haiku | Skill improvement research |
106
107## Priority Order (Reference)
108
109| Category | Priority | Examples |
110| --------------- | ----------- | -------------------------------------- |
111| `architectural` | 1 (Highest) | State stores, API design, core systems |
112| `integration` | 2 | API integration, third-party services |
113| `functional` | 3 | Gameplay mechanics, features |
114| `visual` | 4 | 3D models, materials, textures |
115| `shader` | 4 | Shaders, visual effects |
116| `polish` | 5 (Lowest) | UI styling, visual refinement |
117
118## Category to Agent Mapping (Reference)
119
120| Category | Default Agent |
121| --------------- | ------------- |
122| `architectural` | developer |
123| `functional` | developer |
124| `integration` | developer |
125| `visual` | techartist |
126| `shader` | techartist |
127| `polish` | techartist |
128
129## Usage
130
131Load this router at PM startup:
132
133```
134Skill("pm-router")
135```
136
137Then use `pm-workflow` for the full orchestration.
138
139## References
140
141- [pm-workflow](../pm-workflow/SKILL.md) - Full PM workflow
142- [shared-core](../shared-core/SKILL.md) - Core orchestration concepts
143- [shared-messaging](../shared-messaging/SKILL.md) - Event-driven messaging
majiayu000