agentleFS
DATA-AI-XYZDATA-AI-XYZ/Tandem/skills/draft-epic/SKILL.md

draft-epic

Skill0 starsChanged 6 months agoMITView on GitHub

Draft a new EPIC from an OKR key result or a PRD section. Use when the user asks to draft an epic, create an epic, write an epic, turn an OKR / PRD into an epic, or invokes /tandem:draft-epic. Operates as PM hat. Reads the source OKR or PRD plus the EPIC template; produces 30-Epics/EPIC-NN-<slug>.md with mandatory strategic linkage (frontmatter okr: or prd_section:). Aborts if strategic linkage cannot be established.

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 5 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Follow that prompt verbatim.

    line 16
  2. Distinctive:

    If the user didn't supply one, ask: "Which OKR or PRD am I drafting from?

    line 21
  3. Very specific:

    Use Read / Glob to detect existence.

    line 25
  4. Distinctive:

    Show the file tree of what you'll create before writing: Wait for user approval.

    line 45
  5. Distinctive:

    If the Business Outcome line can't be written ("what metric does this move and by how much?"), stop and ask.

    line 49
  6. Distinctive:

    If the Epic feels > 4 weeks of solo work, propose splitting into two Epics.

    line 50

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “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

Tandem: draft-epic (PM hat)

Operate as PM hat. The user has an approved strategic bet (OKR KR or PRD section) and needs to put it into the work graph as an Epic.

This skill is the slash-command wrapper for the kit's canonical epic-drafting prompt. The prompt is the source of truth; this file is the entry point.

Source of truth

@_00-Project-Management/92-Prompts/02-draft-epic-from-okr-or-prd.md

Follow that prompt verbatim. The sections below add only the slash-command-specific glue — do not re-declare the prompt's content here.

Inputs needed

  • Path to the source OKR file (_00-Project-Management/00-Strategy/OKR-YYYY-Qx.md) or PRD file (_00-Project-Management/20-Requirements/PRD-*.md).
  • If the user didn't supply one, ask: "Which OKR or PRD am I drafting from? Paste the path or the KR/section text inline."

Load into context

Use Read / Glob to detect existence. Treat missing files as "not present" rather than throwing.

  • Source OKR / PRD — at the resolved path above. If neither is provided and no okr: or prd_section: can be inferred from the user's inline text, abort — the kit's strategic-linkage rule forbids Epics without it. Tell the user: "I need an OKR or PRD reference before drafting. Which one moves this?"
  • SOP — _00-Project-Management/90-Standards/SOP.md for DoR, estimation, status enum.
  • Project context — _00-Project-Management/90-Standards/PROJECT-CONTEXT.md.
  • EPIC template — _00-Project-Management/91-Templates/EPIC.template.md. Use verbatim — do not redraft section headings from memory.
  • Existing Epics — glob _00-Project-Management/30-Epics/EPIC-*.md to find the next-free EPIC-NN (scan for the max NN, increment by 1, pad to 2 digits). Do not invent a number; do not reuse one.
  • Project root CLAUDE.md — for project-specific overrides.

Task

  1. Find next-free EPIC-NN by globbing _00-Project-Management/30-Epics/EPIC-*.md and computing max(NN) + 1. Pad to 2 digits.
  2. Draft the Epic at _00-Project-Management/30-Epics/EPIC-NN-<slug>.md using EPIC.template.md verbatim.
  3. Fill every section, especially:
    • Strategic linkage — must reference the source OKR KR or PRD section. Frontmatter okr: OR prd_section: is mandatory — if neither can be set with a real value, abort and tell the user the linkage is too weak to proceed (kit's strategy-linkage rule).
    • In scope / Out of scope — explicit deferrals are how mid-epic creep is resisted.
    • Success criteria — measurable, not vibes.
    • Dependencies, Data touched, Risks — top 3 risks max, with one-line mitigations.
  4. Outline 3–7 Features at high level (title + one-line goal each) in the ## Features section. Do not create FEAT files — /tandem:split-into-features does that.
  5. Set frontmatter: status: not-started, created_at: <ISO 8601 now from system clock>, other timestamp fields empty strings.
  6. Show the file tree of what you'll create before writing. Wait for user approval.

Output rules

  • If the Business Outcome line can't be written ("what metric does this move and by how much?"), stop and ask. Don't proceed without a clear answer.
  • If the Epic feels > 4 weeks of solo work, propose splitting into two Epics.
  • Strategic-linkage rejection: if the user pushes back on the OKR/PRD requirement, point to _00-Project-Management/90-Standards/SOP.md ("Strategy linkage" rule). Do not write the file without okr: or prd_section: set.

Non-negotiable rules from CLAUDE.md

  • Frontmatter timestamps (quoted ISO 8601 with offset, from system clock).
  • Status enum (not-started on creation).
  • Strategy-linkage rule — okr: or prd_section: is MANDATORY. Reject the request and stop if neither can be set.
  • Templates rule — use EPIC.template.md verbatim.

End-of-session summary (always emit)

  • File written: _00-Project-Management/30-Epics/EPIC-NN-<slug>.md
  • Strategic linkage: okr: <ref> or prd_section: <ref>
  • Features outlined: X
  • Estimate: S | M | L | XL
  • Status: not-started

Next command

Next: /tandem:split-into-features — decompose this Epic into FEAT files.

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.