agentleFS
Sign inStart free
mandarzopemandarzope/fullstack-swarm/skills/ship-prd/SKILL.md

ship-prd

Skill4 starsChanged 4 months agoMITView on GitHub

Kick off the agent swarm pipeline from prd/current.md. Reads the PRD, dispatches project-manager to seed _queue.json with track-tagged stories, then runs the autonomous pipeline (parallel batches, same as /next, including UX design and visual QA for ui/fullstack stories) until BLOCKED, all DONE, or stuck. Use when the user types /ship-prd or says "start the pipeline", "kick off the swarm", "process the PRD", "run autonomously from the PRD".

The file

We split this skill into the rules it gives an agent

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

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    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. Distinctive:

    Read CLAUDE.md to refresh the orchestration rules and track routing.

    line 13
  2. Very specific:

    If prd/current.md is missing, tell the user and stop.

    line 18
  3. Very specific:

    If _queue.json is non-empty, ask the user whether to overwrite (start fresh) or merge (decompose only the new sections of the PRD).

    line 20
  4. Very specific:

    If both conditions are clean, dispatch the project-manager subagent at xhigh effort (Task model: claude-opus-4-8[effort=xhigh] or …

    line 23
  5. Very specific:

    After the project-manager returns, parse the <status> block, update stories/_queue.json accordingly, and announce the queue state.

    line 34
  6. Very specific:

    Do not wait for a separate /next between stages.

    line 41

≈ “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

Read CLAUDE.md to refresh the orchestration rules and track routing. Then verify:

  1. prd/current.md exists and is non-empty.
  2. stories/_queue.json exists and contains an empty stories array.

If prd/current.md is missing, tell the user and stop.

If _queue.json is non-empty, ask the user whether to overwrite (start fresh) or merge (decompose only the new sections of the PRD). Do not proceed without an answer.

If both conditions are clean, dispatch the project-manager subagent at xhigh effort (Task model: claude-opus-4-8[effort=xhigh] or account-equivalent). Reinject identity:

Identity: role_id=project-manager, marker=PM. Non-negotiables: do not invent PRD answers; tag every story backend|ui|fullstack; touch only stories/.

Decompose prd/current.md into user stories, tagging each with a track (backend/ui/fullstack). Read stories/_schema.json before writing stories/_queue.json. Follow your subagent definition exactly. Surface any ambiguities as questions_for_human. End with the mandatory <status> block (notes must start with PM:).

After the project-manager returns, parse the <status> block, update stories/_queue.json accordingly, and announce the queue state.

Autonomous continuation: Immediately run the same autonomous pipeline loop as the next skill: keep dispatching in priority order with parallel batches (cap 7), track-routing to the right implementer and through UX design / visual QA for ui/fullstack stories, updating _queue.json after each batch, until BLOCKED, all DONE, or no dispatchable work. Do not wait for a separate /next between stages.

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.

Check your setup, free