agentleFS
Sign inStart free
bex-cobex-co/bex-security/.claude/skills/loop-worker/SKILL.md

loop-worker

Skill54 starsChanged 3 months agoApache-2.0View on GitHub

Autonomously work a `.pm` workstream to completion — pick the next pending milestone, implement every task end to end, /ship it, then move to the next until none remain. Use when the user asks to "loop", drain, or work through a whole workstream's backlog (e.g. `/loop-worker w1`). Sequential, not interval-based; for a timed poll use /loop.

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 1

    1. jpoindexter/theft-kit★ 1same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository8 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

10 rules. About 0 are common, 5 are distinctive, and 5 are 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.

  • 5

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  • could be automatic: a setting could enforce it
  1. Distinctive:

    If there are no pending milestones, go to Exit.

    line 26
  2. Distinctive:

    Do the actual engineering, task by task, in the order the milestone implies:

    line 36
  3. Very specific:

    Follow all AGENTS.md rules (root AGENTS.md and sdk/typescript/AGENTS.md — keep it simple, no speculative defenses, public CLI surface …

    line 38
  4. Very specific:

    Run the relevant test suites and make them pass before considering a task done — from sdk/typescript/: pnpm run test, pnpm run types, pnpm …

    line 39
  5. Very specific:

    If /ship surfaces a failure it cannot fix (rebase conflict it can't resolve, rejected push), treat it as a block (see below).

    line 47
  6. Distinctive:

    Stop the loop and give a final summary when any of these holds:

    line 55
  7. Very specific:

    Done: no pending milestones remain in <wN>.

    line 57
  8. Distinctive:

    One milestone per ship: Never batch two milestones into one commit.

    line 63
  9. Very specific:

    Stay in <wN>: Only pick milestones from the requested workstream.

    line 65
  10. Distinctive:

    Report honestly: If you skipped a task, mocked something, or a suite was flaky, say so in the per-milestone summary — don't present partial …

    line 66

≈ “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

Task: Drain a .pm workstream milestone by milestone

/loop-worker <wN> — repeatedly pick the next pending milestone in workstream <wN>, implement it fully, /ship it, and continue to the next, until the workstream has no pending milestones left (or a milestone genuinely blocks). This is a long-running autonomous loop over the .pm board; it composes /pm (board reads/writes), your own implementation work, and /ship.

Parse the target workstream from $ARGUMENTS (e.g. w1). If $ARGUMENTS is empty, STOP and ask which workstream to drain — never guess.

Preconditions (verify once, up front)

  1. git branch --show-current is main. If not, STOP and ask (same rule as /ship).
  2. git status — note pre-existing uncommitted changes. Do not sweep unrelated changes into a milestone's ship; if the tree is dirty with work you didn't do, surface it and ask before starting.
  3. The workstream .pm/<wN>/README.md exists. If not, STOP and report.

The loop

Repeat until the exit condition below:

1. Pick the next pending milestone

Read .pm/<wN>/README.md. In the ## Milestones list, pending milestones are the unchecked ones (- [ ] **mN**). Pick the lowest-numbered pending milestone that still has a live directory (.pm/<wN>/mN/, not under done/). Cross-check by listing .pm/<wN>/m*/ and its done/ folder — the checkbox and the on-disk state must agree; if they disagree, trust the task files and flag the drift.

If there are no pending milestones, go to Exit.

Announce which milestone you picked and give a one-line plan before doing work.

2. Understand the milestone

Read .pm/<wN>/mN/README.md and every task file .pm/<wN>/mN/tNNN.md (skip any already in done/). These define the scope, order, and acceptance. Milestones ship features end to end — include the SDK, docs, and container tasks alongside the CLI ones; do not stop at the code. Respect .pm/DO_NOT_DO.md and the milestone's own acceptance criteria.

3. Implement it

Do the actual engineering, task by task, in the order the milestone implies:

  • Follow all AGENTS.md rules (root AGENTS.md and sdk/typescript/AGENTS.md — keep it simple, no speculative defenses, public CLI surface discipline, public-repository hygiene; prettier on .pm markdown, skill layout, etc.).
  • Run the relevant test suites and make them pass before considering a task done — from sdk/typescript/: pnpm run test, pnpm run types, pnpm run format (see sdk/typescript/TESTING.md), plus a container build when Dockerfile/docker/ change — whichever the change touches. Never mark a task complete on unverified code.
  • You may delegate independent sub-tasks to subagents (Agent tool) to parallelize, but you own correctness.
  • Keep the milestone's .pm status in sync as you finish tasks (task frontmatter, milestone README **Status:** + the — DONE row, workstream checkbox) — these are /pm-governed writes, so follow the conventions in .claude/skills/pm/SKILL.md exactly. When the milestone is fully done, move it per the .pm rule: a done milestone with no open tasks moves whole to .pm/<wN>/done/mN/ (mv the directory, then rmdir the empty original — leave no tombstone), and its workstream checkbox flips to [x].

4. Ship it

Invoke /ship (the ship skill) for this milestone's changes. Because you made the changes this session, /ship runs session-aware: it stages exactly what you touched and writes the commit message from your knowledge. /ship ends at a successful push — it does not watch CI or the deploy. Do not proceed to the next milestone until /ship reports the shipped HEAD.

If /ship surfaces a failure it cannot fix (rebase conflict it can't resolve, rejected push), treat it as a block (see below).

5. Continue

Loop back to step 1 to pick the next pending milestone.

Exit

Stop the loop and give a final summary when any of these holds:

  • Done: no pending milestones remain in <wN>. Report which milestones you shipped this run.
  • Blocked: a milestone needs a decision only the user can make (ambiguous scope, a DO_NOT_DO conflict, an external credential/access you lack, or a ship failure you can't resolve). Stop before shipping half-work — report the exact blocker and what you'd need to proceed. Do not skip a blocked milestone to grab a later one unless the user says so.
  • Budget/interrupt: the user interrupts, or you've been running long enough that a checkpoint is warranted — report progress (shipped, in-flight, remaining) so the run can be resumed cleanly.

Guardrails

  • One milestone per ship. Never batch two milestones into one commit; each milestone lands as its own shipped unit so history and rollback stay clean.
  • Never ship red. A failing test suite is a block, not a footnote — don't route around it.
  • Stay in <wN>. Only pick milestones from the requested workstream. Workers are general-purpose, but this run is scoped to the queue the user named.
  • Report honestly. If you skipped a task, mocked something, or a suite was flaky, say so in the per-milestone summary — don't present partial work as complete.

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