Skip to content
shell
$ npx -y skills add aleksander-dytko/ai-pm-workspace --skill create-epic --agent claude-code

How it fires

How this skill gets triggered: by you, by Claude, or both.

  • Fires itselfAuto-invocation. Claude auto-loads it when your prompt matches the work.
  • You can call itInvoke it directly when you want it.
  • Slash command/create-epic
How auto-invocation works

Context preview

The summary Claude sees to decide when to auto-load this skill.

Turn an idea into a structured epic draft using templates/epic-template.md

SKILL.md

create-epic.SKILL.md
name: create-epic
description: Turn an idea into a structured epic draft using templates/epic-template.md
argument-hint: "[epic idea - can be a short prompt or a path to a raw idea note]"

Create Epic

You help turn a raw product idea into a well-structured epic draft. The output follows `templates/epic-template.md` and lands in `epics/` with `stage: Explore`.

Input

The user provides via `$ARGUMENTS`:

  • A short prompt (e.g., "unified onboarding dashboard for CSMs"), OR
  • A path to a raw idea note (e.g., `samples/sample-epic-idea.md`)

Workflow

1. Gather context

**Read the raw idea.** If `$ARGUMENTS` is a path, read that file. Otherwise use the prompt directly.

**Read supporting context in parallel**:

  • `templates/epic-template.md` - the target structure
  • `docs/epic-lifecycle.md` - the stage framework
  • `research/personas/` - personas that might match the target user
  • `Dashboard/people-profiles.md` - any stakeholders mentioned
  • `Loose Notes/Work/` and `Meetings/` - related prior discussions (search by keywords from the idea)
  • `Initiatives/` - does this fit under an existing initiative?

2. Identify gaps in the idea

Before drafting, identify what's missing or unclear. Ask the user 3-5 focused questions in ONE `AskUserQuestion` round if any of these are unclear from the raw input:

  • **Target user**: which persona does this serve? (Offer personas from `research/personas/` as options.)
  • **Primary metric**: what measurable outcome should improve?
  • **Scope boundary**: what is explicitly NOT in this epic?
  • **Initiative fit**: does this belong under an existing initiative (show list from `Initiatives/`) or does it stand alone?
  • **Competitor check**: has competitive research been done? (Offer `/competitive-research` if not.)

Skip questions where the raw idea already has a clear answer. Maximum one round.

3. Draft the epic

Use `templates/epic-template.md` as the structure. Fill what you can from the raw idea + answers:

**Required sections (always fill)**:

  • **Value Proposition Statement** - one crisp statement of the outcome.
  • **User Problem** - 1-2 paragraphs from the customer's perspective. Frame from the "why", not the product's failure.
  • **User Stories** - 3-6 stories at the "what" level (Explore phase). Add the "how" later.

**Phase-specific (fill lightly in Explore, flag for Scope)**:

  • **Breakdown** - list sub-activities under Explore (customer interviews, research, opportunity-solution tree).
  • **Risk Assessment** - initial class only, full in Scope.

**Skip for now**:

  • Release Notes (drafted in Scope or later).
  • Implementation Notes (Scope phase).
  • Design Planning, Documentation Planning (Scope phase).
  • Validation Criteria (Validate phase).

**Fill placeholders honestly.** If you don't know something, write `[To determine during Scope]` or `[Needs user research]`. Don't fabricate.

4. File location and naming

  • **Folder**: `epics/`. Create if it doesn't exist.
  • **Filename**: `YYYY-MM-DD - [Epic Title slug].md`. Example: `2026-04-20 - Unified onboarding dashboard.md`.
  • **Frontmatter**: set `stage: Explore`, `status: exploring`, `created: YYYY-MM-DD`.

5. Link the epic

  • **Daily journal**: add a link in today's journal under `## Notes`. Format: `- [[YYYY-MM-DD - Epic Title]] - new epic in Explore stage`.
  • **Initiative**: if the epic fits under an existing initiative (from step 2), add a link in the initiative file under `## Epics under this initiative`.
  • **Nothing external**: do NOT push to any external tracker. The epic lives in the vault until you explicitly ship it elsewhere.

6. Propose follow-up tasks

Based on what's missing in the draft, propose 2-5 tasks that would move the epic out of Explore. Examples:

  • Schedule 5 customer interviews with target persona
  • Run `/competitive-research` on [topic]
  • Run `/opportunity-solution-tree` for the primary metric
  • Review the draft with [stakeholder]
  • Draft a 3-bullet research brief for the user research team

Show the proposed task list and wait for user confirmation. After confirmation, append to `Dashboard/tasks.md` under "This week" with format:

- [ ] Task - source: [[YYYY-MM-DD - Epic Title]] - due: YYYY-MM-DD - priority: P3

7. Output summary

✅ Epic drafted: [[YYYY-MM-DD - Epic Title]]
✅ Linked in today's journal
[if initiative match] ✅ Linked under [[Initiative]]

Stage: Explore
Status: exploring

📋 Proposed follow-up tasks (to move to Scope):
1. [Task]
2. [Task]

Confirm: Add tasks to Dashboard/tasks.md? (Y/N)

💡 Next steps in the lifecycle: see docs/epic-lifecycle.md

Notes

  • **Honesty over completeness.** It's better to mark a section as `[Needs user research]` than to invent answers.
  • **Epics start thin and grow.** The Explore-stage draft should be under 1 page. Sections get filled as the epic progresses.
  • **One initiative per epic.** If the idea fits under multiple initiatives, flag this as a scope problem rather than dual-linking.
  • **User stories from the customer's perspective**, not the product's. "As a support engineer, I can..." not "The product shall...".
Read more
Read it on GitHub ↗
Ships withai-pm-workspace

A structured, Claude-Code-ready workspace for product managers, designers, and other PM-adjacent roles. Your notes become the persistent memory; Claude Code becomes the executor. This is a template.

Get the whole plugin, auto-invoked
Stats
32
Stars
0
Views
5
Forks
Maintained
Maintenance
MIT
License
3mo ago
Last commit
5mo ago
Created

Repo: aleksander-dytko/ai-pm-workspace