Skill Claude

PM: Scope-Disciplined Product Manager

pm/

2 views 0 downloads

A product manager for your agent team: build-ready PRDs with testable gates, a roadmap, brainstorming, standups and a single launch or no-launch verdict, with a scope guard that cites the spec.

PM: Scope-Disciplined Product Manager: Brain Lab skill for Claude Code

The scope-disciplined product manager: plans the project, gates the launch, and says no.

Why use it

Requirements that can't be tested, "the flow should feel smooth" is not a gate.
Scope creep accepted one small ad-hoc ask at a time.
"Looks ready" as a launch decision, with no scenario actually scored.

What it does

prd, turn the spec into a build-ready PRD: objectives, scope, out-of-scope guards, every success scenario mapped to a testable gate, the build plan.
roadmap, own the docs/ roadmap: epics, user stories decomposed across the COMPLETE user-flow graph, milestones, and a machine-readable mirror.
brainstorm, explore an idea with you before anything is specified: one question at a time, 2-3 approaches, a design you approve.
acceptance, the terminal gate: every scenario PASS/FAIL with evidence, 100% end-to-end breadth, ONE binary LAUNCH / NO-LAUNCH verdict.
standup, evidence-based Done / Next / Blockers, read-only.
scope-guard, runs inside every mode: "is X in scope?" answered Yes/No with the spec section cited.

Benefits

Requirements a machine can score. Testable gates, not adjectives.
Scope creep stops here. Every "no" cites the spec and names the edit that would reverse it.
No half-covered features. Stories map every path a user can take, not one happy path.
Launch calls you can defend: one binary verdict, backed by per-scenario evidence.
Retarget it in one file. Nothing about your product is baked in.

Install

Download this skill and copy the pm folder to .claude/skills/pm/ in your project, or to ~/.claude/skills/pm/ to use it in every project. Claude Code finds it on its own; you can also call it with /pm.

Part of the Agent Engineering bundle. To get the whole set as a Claude Code plugin: /plugin marketplace add endorphin-ai/hasbrains-agent-kit, then /plugin install agent-engineering@hasbrains-agent-kit.

Source: endorphin-ai/hasbrains-agent-kit. License: MIT.

Files (19)

{
  "skill": "pm",
  "version": "1.0",
  "kind": "second-brain",
  "entry": "SKILL.md",
  "maps": { "index": "maps/index.json", "tags": "maps/tags.json", "links": "maps/links.json" },
  "companions": ["FORMAT.md", "VOICE.md"],
  "workflows": [
    "workflows/prd.md",
    "workflows/roadmap.md",
    "workflows/brainstorm.md",
    "workflows/acceptance.md",
    "workflows/standup.md",
    "workflows/scope-guard.md"
  ],
  "references": ["references/roadmap-json.md", "references/standup-data-sources.md"],
  "note_count": 8,
  "required_companion_skill": {
    "name": "docs-project-management",
    "why": "Defines every file this PM writes — work-item taxonomy, link graph, status lifecycle, frontmatter fields, test-case template, bug + RCA format, team work-reporting.",
    "url": "https://github.com/endorphin-ai/hasbrains-agent-kit/tree/main/agent-engineering/skills/docs-project-management"
  },
  "optional_companion_skills": ["brainstorming", "build-brain", "pipeline-state", "four-principles", "verification-before-completion", "ux-expert"],
  "project_facts_source": "the project's own config + spec (this kit's convention: docs/project_config/info.md + a spec file)",
  "modes": ["prd", "roadmap", "brainstorm", "acceptance", "standup"],
  "continuous": ["scope-guard"]
}