agentleFS
cenconq25cenconq25/claude-code-app-studio/.claude/skills/create-epics/SKILL.md

create-epics

Skill40 starsChanged 7 months agoMITView on GitHub

Translates approved PRDs and architecture into epics — one epic per architectural module or major system. Each epic captures scope, governing ADRs, framework risk, and untraced requirements. Run after /architecture-review passes; do NOT decompose into stories — run /create-stories per epic afterward.

10 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use vertical-slice.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • hongyuancSkill

    vertical-slice

    Build and evaluate a production-quality vertical slice before advancing from Pre-Production to Production.

    Also in 4 projects that use this

The file

We split this skill into the rules it gives an agent

12 rules. 0 are common, 7 are distinctive, and 5 are very specific to this project.

  • 0

    Common

    Most of the 10 edited versions we read keep these.

  • 7

    Distinctive

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Do not decompose into stories here — /create-stories <epic-slug> does that.

    line 25
  2. Very specific:

    If a previous epic for the same system exists, ask: Refresh, Replace, Cancel

    line 46
  3. Very specific:

    If <system-name> argument given, scope to that system.

    line 48
  4. Distinctive:

    Read the inputs in full. Hold:

    line 54
  5. Distinctive:

    If a system is huge (>30 TR-IDs), split into 2-3 sub-epics (e.g., "Auth — Sign-in", "Auth — Account & Recovery")

    line 68
  6. Distinctive:

    After all epics are written:

    line 184
  7. Very specific:

    Update design/systems-index.md to set Status to Epic created for each affected system

    line 185
  8. Very specific:

    Update production/session-state/active.md with epic count and next-action pointer

    line 186
  9. Distinctive:

    If any epic has Section 6 entries (untraced requirements), warn loudly:

    line 204
  10. Distinctive:

    Every TR-ID from every in-scope PRD appears in either Section 5 (covered) or Section 6 (untraced) of some epic

    line 211
  11. Distinctive:

    Every Governing ADR listed has Status verified — Proposed ADRs trigger BLOCKED labeling

    line 212
  12. Distinctive:

    No epic has zero TR-IDs (else it is not a real epic — ask the user)

    line 214

“Common” is measured: kept by at least 6 in 10 of the 10 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Create Epics

An epic is the unit between architecture and stories. It groups related stories under a shared system or module, declares the governing ADRs, and surfaces the framework risk so the engineer knows what to verify before coding.

Output: production/epics/<epic-slug>/EPIC.md (one file per epic)


Purpose / When to Run

Run when:

  • /architecture-review returned PASS or CONCERNS (not FAIL)
  • /create-control-manifest is written and dated
  • The team is ready to plan stories

Do not decompose into stories here — /create-stories <epic-slug> does that.

Inputs

  • All design/prd/*.md
  • All Accepted docs/architecture/ADR-*.md
  • docs/architecture/architecture.md
  • docs/architecture/control-manifest.md
  • docs/framework-reference/<framework>/
  • Optional <system-name> argument: only generate the epic for this system

Outputs

  • production/epics/<epic-slug>/EPIC.md — one per system in scope

Phase 1: Pre-checks

  • Confirm at least one Accepted ADR exists
  • Confirm docs/architecture/control-manifest.md exists and read its Manifest Version
  • If a previous epic for the same system exists, ask: Refresh, Replace, Cancel

If <system-name> argument given, scope to that system; else, do all in-scope MVP systems from the systems index.


Phase 2: Read All Context

Read the inputs in full. Hold:

  • Each PRD's Detailed Requirements (all TR-IDs)
  • Each PRD's Acceptance Criteria
  • Each ADR's PRD Requirements Addressed table
  • Architecture Section 3 (Layers) and Section 20 (Required ADRs)
  • Control manifest Required/Forbidden/Guardrail by layer

Phase 3: Group TR-IDs into Epics

The mapping rule: one epic per system, with stories living inside.

Special cases:

  • If a system is huge (>30 TR-IDs), split into 2-3 sub-epics (e.g., "Auth — Sign-in", "Auth — Account & Recovery")
  • Foundation systems (state, navigation, networking) become epics if they need any user-visible scaffolding work
  • Cross-cutting concerns (analytics, error reporting) become their own epics if they have ≥3 TR-IDs of dedicated work

Order epics by dependency: Foundation → Core → Features → Polish.


Phase 4: Per-Epic Authoring

For each epic, write production/epics/<slug>/EPIC.md:

# Epic: <System Name>

> **Status**: Ready for stories
> **Layer**: <Foundation / Core / Feature / Polish>
> **Manifest Version**: <date from control-manifest>
> **Created**: <date>

## 1. Goal
<one paragraph from the PRD's Overview + User Goal>

## 2. PRD Source
- `design/prd/<system>.md`

## 3. Governing ADRs

| ADR | Title | Status | Risk |
|-----|-------|--------|------|
| ADR-NNNN | <title> | Accepted | LOW / MED / HIGH |
| ADR-NNNN | <title> | Proposed | (BLOCKED until Accepted) |

## 4. Framework Risk

- Framework: <name + version>
- Knowledge risk: <LOW / MEDIUM / HIGH>
- Verification required (from ADRs): <list>
- Breaking-change touch points: <list or none>

## 5. Requirements Covered

| TR-ID | Requirement summary | Source ADR | Story slug (when created) |
|-------|---------------------|------------|---------------------------|
| TR-<sys>-001 | ... | ADR-NNNN | (pending) |
| ... | | | |

## 6. Untraced Requirements

| TR-ID | Why untraced | Resolution path |
|-------|--------------|-----------------|
| TR-<sys>-007 | No ADR covers this — Proposed only | Run `/architecture-decision` for <area> |

If empty, mark "None — all TR-IDs trace to an Accepted ADR."

## 7. Acceptance for Epic Done

- All TR-IDs in Section 5 implemented and verified
- All stories in this epic at Status: Complete
- All test evidence at the path specified in each story's manifest
- Smoke check passes for this system
- All P0 ADRs covering this epic remain Accepted (no regressions)

## 8. Out of Scope

- <items deliberately excluded — link to where they live (next epic, deferred PRD, v1.1 list)>

## 9. Implementation Notes

### Layer constraints
<from control manifest, copied verbatim — Required and Forbidden for each layer this epic touches>

### Performance budget
<from control manifest guardrails relevant to this epic>

### Test evidence

| Story type expected | Evidence required | Path |
|---------------------|-------------------|------|
| Logic | Automated unit test | tests/unit/<system>/<slug>_test.<ext> |
| Integration | Integration test or documented walkthrough | tests/integration/<system>/<slug>_test.<ext> |
| Visual / Feel | Screenshot + lead sign-off | production/qa/evidence/<slug>-evidence.md |
| UI | Manual walkthrough doc | production/qa/evidence/<slug>-evidence.md |
| Config / Data | Smoke check | production/qa/smoke-<date>.md |

## 10. Risks

- <list — engineering risks that could derail this epic>

## 11. Rollout & Feature Flag

- Feature flag controlling this epic (if any): `<flag_name>` from <provider>
- Rollout plan: <internal → beta → percentage → full>
- Kill switch: <yes / no>

## 12. Open Questions

- <list>

## 13. Story Backlog (filled by `/create-stories`)

[Initially empty. After `/create-stories <slug>`, this lists the story files.]

Phase 5: Write & Confirm

Show the draft for the first epic. Ask: "Approve and write to production/epics/<slug>/EPIC.md?"

Loop for each in-scope epic. The user can approve all in batch or one-by-one.


Phase 6: Update Indexes

After all epics are written:

  • Update design/systems-index.md to set Status to Epic created for each affected system
  • Update production/session-state/active.md with epic count and next-action pointer

Phase 7: Hand Off

Print:

"<N> epics written. Run /create-stories <epic-slug> for each in dependency order. Foundation epics first."

Use AskUserQuestion:

  • Run /create-stories for <first foundation epic>
  • Run /sprint-plan (after at least one story batch is created)
  • Stop here

Phase 8: Detect Untraced Requirements

If any epic has Section 6 entries (untraced requirements), warn loudly:

"<N> requirements across <M> epics are untraced. Stories cannot be implemented safely against untraced requirements. Resolve by writing missing ADRs (/architecture-decision) or marking the requirement out of scope for v1."


Quality Gates

  • Every TR-ID from every in-scope PRD appears in either Section 5 (covered) or Section 6 (untraced) of some epic
  • Every Governing ADR listed has Status verified — Proposed ADRs trigger BLOCKED labeling
  • Manifest Version is recorded at epic creation time
  • No epic has zero TR-IDs (else it is not a real epic — ask the user)

Examples

/create-epics

  • 8 in-scope systems → 9 epics (auth split into Sign-in and Account)
  • 47 TR-IDs covered, 0 untraced
  • Foundation epics: navigation-shell, auth-sign-in, persistence
  • Output: 9 EPIC.md files

/create-epics auth

  • Single-system run. 12 TR-IDs covered by ADR-0003 + ADR-0007. 1 untraced (biometric unlock — no ADR yet).
  • Warning surfaces; user chooses to write ADR before story creation.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.