Skip to content

ndisisnd/msg / .claude/skills/plan-pm

plan-pm

Verified

Principal PM skill — the autonomous PRD writer. Consumes a graded, fleshed-out row from the INTAKE.md backlog (idea, goal, type, grade) and drafts the full PRD solo — feature/acceptance table, edge cases, error handling — saved to features/planned/prd-[n]-[feature-slug]/. The requirements interview lives in /intake now, not here. Treats the row as settled and picks its own few follow-up questions (refs/principles.md); pauses ONLY for that one batch and for breaking/critical touches. --update edits an existing PRD's content and syncs its intake row; --migrate converts old PRDs to the current template. --review [on|off] runs plan-review at the end (this run, or saved per project). Refuses requests that would skip the PRD stage entirely.

FileKindSize
SKILL.mdskill md12.3 KB
refs/error-checklist.mdmarkdown2.8 KB
refs/principles.mdmarkdown2.4 KB
refs/protocol-migrate.mdmarkdown8.1 KB
refs/protocol-pm.mdmarkdown17.4 KB
refs/protocol-sub.mdmarkdown6.7 KB
refs/protocol-update.mdmarkdown6.9 KB
refs/template-prd.mdmarkdown11.8 KB

refs/principles.md

View on GitHub
---
name: plan-pm Principles
description: The rules of thumb plan-pm uses to decide which follow-up questions shape and scope a PRD — keep faith with the intake row, flesh out only what the row lacks, never re-ask what intake settled. The only rule for asking in plan-pm.
type: reference
---

# plan-pm Principles

plan-pm drafts solo. These rules decide the few moments it asks the user anything. They are separate from intake's `principles.md`: intake decides *what* the idea is; plan-pm decides *how much detail* the PRD needs. Heuristics, not a checklist.

1. **The row is settled.** The row's `idea`, `goal`, `type` and `grade` hold what intake inferred and asked. Treat every one of them as answered. Never ask a question the row already answers, in any wording.
2. **Never contradict or widen the row.** Every feature, criterion and error case must serve the row's idea and goal. A draft that would change the scope — a new outcome, a new surface, a different user — stops: tell the user and send them to `/intake --update #<n>` to change the row first. The row stays the single source of truth until the PRD exists.
3. **Flesh out, do not reinvent.** The PRD adds what the row lacks: the feature split, the acceptance criteria, the error and edge cases. Fill these from the row, the devkit context and prior PRDs first.
4. **Ask only when two readings give two different PRDs.** A question is worth asking when the answer changes a feature, a criterion, or what is out of scope. A gap that changes nothing is not a question; a gap you can fill from context is not a question.
5. **Pick the questions yourself, and few of them.** Choose the 1–4 questions that most shape the PRD. Offer your best reading as the first option, marked recommended. Batch them into one `AskUserQuestion` (Step 4). Everything else stays a §5 open question or a `[USER: …]` placeholder.
6. **One problem, one PRD.** When the row turns out to hold two user problems, do not draft both. Stop and send the user to `/intake --update #<n>` to split the row.
7. **Never ask for a number you would have to invent.** A time limit, count or threshold either comes from a source or becomes a `[USER: …]` placeholder with a §5 row. Ask about it in Step 4 only when the criterion cannot be tested without it.
8. **The safety pause is not a question budget.** A breaking change, a data change or a production-config touch always pauses (Step 4), on top of these rules.
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 469kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 321kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 321kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 321kUpdated Claude CodeCodexVerifiedInstall