Back to skills
extension
Category: AI Agent CapabilitiesNo API key required

pm-router

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.

personAuthor: jakexiaohubgithub

PM Skill Router

"Right skill for the right phase at the right time."


Route by Current Status (Post-Completion Flow)

⚠️ CRITICAL: When in post-completion workflow, follow this routing table to continue through ALL phases!

| Current Status | Action | Next Skill/Phase | | --------------------------------- | ------------------------------------------------------------ | ------------------------------------ | | retrospective_synthesized | Check if playtest required → Use playtest session skill OR skip to PRD reorganization | pm-retrospective-playtest-session or pm-organization-prd-reorganization | | playtest_complete | Continue to PRD reorganization | pm-organization-prd-reorganization | | playtest_skipped | Continue to PRD reorganization | pm-organization-prd-reorganization | | prd_refinement | Continue to cleanup phase | Cleanup (inline in pm-workflow) | | cleanup_completed | Continue to skill research | pm-improvement-skill-research | | skill_updates_applied | Continue to task selection | pm-organization-task-selection | | task_ready (post-completion) | Create test plan | pm-planning-test-planning | | test_plan_ready | Assign task, send message, EXIT | (Exit - worker has task) |

Playtest Decision (at retrospective_synthesized):

Skip playtest for:

  • Test infrastructure bugfixes
  • Non-gameplay tasks (CI/CD, tooling, documentation)
  • Backend-only changes without visual impact
  • Small typo/minor text fixes

Use playtest for:

  • Gameplay mechanics (movement, shooting, physics)
  • Visual features (shaders, materials, effects)
  • UI/UX changes (HUD, menus, interactions)
  • Character/weapon behavior changes
  • Multiplayer features

Route by Workflow Phase

| Phase | Primary Skills | Sub-Agents | | ---------------------- | -------------------------------------------------------------------- | ------------------------------ | | Startup | pm-organization-scale-adaptive | - | | Task Selection | pm-organization-task-selection, pm-organization-task-research | pm-task-researcher | | Test Planning | pm-planning-test-planning | pm-test-planner | | Assignment | (use selected skills) | - | | Retrospective | pm-retrospective-facilitation, pm-retrospective-playtest-session | pm-retrospective-facilitator | | PRD Update | pm-organization-prd-reorganization | pm-prd-organizer | | Skill Research | pm-improvement-skill-research, pm-improvement-self-improvement | skill-researcher | | Architecture Check | pm-validation-architecture | pm-architecture-validator |

Route by Task Category

| Category | Agent | PM Skills to Load | | -------------------- | ---------- | --------------------------------------------------------------------- | | architectural | developer | pm-organization-task-research, pm-validation-architecture | | integration | developer | pm-organization-task-research | | functional | developer | pm-organization-task-research | | visual | techartist | pm-configuration-asset-coordination, pm-configuration-vite-assets | | shader | techartist | pm-configuration-asset-coordination | | polish | techartist | pm-configuration-asset-coordination | | Any with multiplayer | developer | pm-validation-architecture |

Route by Signal Keywords

| Signal in Task | Route To | | ---------------------------------------------------- | ------------------------------------- | | "asset", "model", "fbx", "texture", "gltf" | pm-configuration-asset-coordination | | "vite", "public/", "src/assets/" | pm-configuration-vite-assets | | "multiplayer", "server", "colyseus", "authoritative" | pm-validation-architecture | | "test", "e2e", "validation", "acceptance" | pm-planning-test-planning | | "gdd", "design", "mechanic" | pm-organization-prd-reorganization |

All PM Skills

| Skill | Purpose | Phase | | ------------------------------------- | ------------------------------------- | -------------- | | pm-organization-scale-adaptive | Adjust planning depth by PRD size | Startup | | pm-organization-task-selection | Priority algorithm for task selection | Task Selection | | pm-organization-task-research | Codebase research before assignment | Task Selection | | pm-planning-test-planning | Collaborative test planning | Test Planning | | pm-configuration-asset-coordination | Asset coordination for parallel work | Assignment | | pm-configuration-vite-assets | Vite 6 asset patterns | Assignment | | pm-validation-architecture | Server-authoritative validation | Assignment | | pm-retrospective-facilitation | Worker retrospective orchestration | Retrospective | | pm-retrospective-playtest-session | Game Designer playtest coordination | Retrospective | | pm-organization-prd-reorganization | GDD-to-PRD task extraction | PRD Update | | pm-improvement-skill-research | Multi-agent skill improvements | Skill Research | | pm-improvement-self-improvement | PM self-improvement | Skill Research |

Sub-Agents (via Task tool)

| Sub-Agent | Model | Purpose | | ------------------------------ | ------- | ---------------------------------------- | | pm-task-researcher | Haiku | Codebase research before task assignment | | pm-test-planner | Inherit | Test planning with QA+GD | | pm-retrospective-facilitator | Inherit | Retrospective orchestration | | pm-prd-organizer | Inherit | PRD reorganization | | pm-architecture-validator | Haiku | Architecture gap detection | | skill-researcher | Haiku | Skill improvement research |

Priority Order (Reference)

| Category | Priority | Examples | | --------------- | ----------- | -------------------------------------- | | architectural | 1 (Highest) | State stores, API design, core systems | | integration | 2 | API integration, third-party services | | functional | 3 | Gameplay mechanics, features | | visual | 4 | 3D models, materials, textures | | shader | 4 | Shaders, visual effects | | polish | 5 (Lowest) | UI styling, visual refinement |

Category to Agent Mapping (Reference)

| Category | Default Agent | | --------------- | ------------- | | architectural | developer | | functional | developer | | integration | developer | | visual | techartist | | shader | techartist | | polish | techartist |

Usage

Load this router at PM startup:

Skill("pm-router")

Then use pm-workflow for the full orchestration.

References