agentleFS
Fighter90Fighter90/career-ops-ui/.agents/skills/swarm/SKILL.md

swarm

Skill99 starsChanged 8 days agoMITView on GitHub

Launch the career-ops-ui agent swarm: tasks from Linear (state Todo + label `swarm`) are carried automatically through develop → gates → review → document → release (PR, CI, version, tag, deploy) → prod-QA, and Linear is updated at every step. Use when the user says «запусти рой», «запусти навык Рой», «run the swarm», «swarm run», or asks to process swarm-labeled Linear tasks.

  • Commits and pushes

98 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use poteto-help, make-bot-ui and 10 more.

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.
  • ianclearySkill

    poteto-help

    Guides users through pstack setup, /poteto-mode, and picking the skill, playbook, or principle for a task. Type /poteto-help with a question.

    Also in 14 projects that use this

  • eryzerzSkill

    make-bot-ui

    Use when building a custom UI (page, dashboard, buttons) that should wake a Grok Bot over a webhook, when the user must provide a webhook sender key, or when exposing that UI on Tailscale.

    Also in 19 projects that use this

  • ianclearySkill

    principle-explain-the-number

    Apply before you trust, report, or act on a number you measured: a speedup, a regression, a throughput, a latency, or an eval result. Find what limits it, and rule out that it measured something other than the work you think.

    Also in 27 projects that use this

  • bnemaSkill

    setup-pstack

    Configure which Pi models pstack uses per role. Detects available provider/model specs and writes pstack's user configuration. Use for /setup-pstack, "configure pstack models", or changing pstack model choices.

    Also in 23 projects that use this

  • jaydubya818Skill

    principle-attack-the-premise

    Apply when repeated fixes fail and a shared assumption may be wrong.

    Also in 31 projects that use this

  • jaydubya818Skill

    poteto-mode

    Use when non-trivial work should follow the smallest rigorous Jstack workflow.

    Also in 27 projects that use this

  • +5 more from jaydubya818

The file

We split this skill into the rules it gives an agent

6 rules. 0 are common, 3 are distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 94 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 3

    Very specific

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

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

    Run it in the background; stream the log to the user (tail -f path) — the user asked to see the trace in real time.

    line 33
  2. Distinctive:

    Every failure mode so far was an orchestrator bug that a regression test now pins (placeholder echo, permission-denied edits, lost branch …

    line 44
  3. Distinctive:

    When a run fails:

    line 46
  4. Very specific:

    Read the latest .task-board/log/*.log and the failed issue's Linear comments.

    line 48
  5. Very specific:

    never run the swarm from a non-main branch.

    line 56
  6. Distinctive:

    When the run ends, summarize to the user in THEIR language: which issues were picked, what shipped (version, PR, CI runs, prod-QA …

    line 63

“Common” is measured: kept by at least 6 in 10 of the 94 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

Swarm run

You are launching the agent swarm of this repository (career-ops-ui). The machinery lives in scripts/swarm/ (orchestrator + role prompts, architecture in scripts/swarm/README.md). Your job is to preflight, launch, babysit, and report — the orchestrator does the rest.

1. Preflight (all must pass)

git -C <repo> status --porcelain        # must be EMPTY
git -C <repo> rev-parse --abbrev-ref HEAD   # must be `main`
opencode --version                      # the worker CLI must exist

<repo> is this repository root. If the tree is dirty, commit or stash FIRST (never launch the swarm on a dirty tree — it self-checks and aborts). If tasks should come from a queue of N, note the user's N (default 1).

2. Launch

node scripts/swarm/orchestrator.mjs --max <N> 2>&1 | tee /tmp/swarm-run-$(date +%H%M).log
  • Flags you may pass through: --max N (tasks this run), --rounds N (dev/review rounds, default 3), --dry-run (pick + plan only), --no-release (stop after review ACCEPT).
  • SWARM_REMOTE_QA_LIVE=1 env adds the 19-locale live LLM leg to the prod QA (expensive).
  • Run it in the background; stream the log to the user (tail -f path) — the user asked to see the trace in real time.

3. Babysit (the swarm talks through its log)

Milestones to relay: picked <ID> → claim → round n: develop → gates green → verdict: ACCEPT|RETURN → story: → merged: → release chain (ok=true version=…) → remote-qa verdicts → === DONE ===. Read .task-board/log/<cycle>.log for worker tails.

4. On failure — fix forward, do not shrug

Every failure mode so far was an orchestrator bug that a regression test now pins (placeholder echo, permission-denied edits, lost branch commits, already-merged work). When a run fails:

  1. Read the latest .task-board/log/*.log and the failed issue's Linear comments.
  2. Classify: orchestrator bug → fix it red-first (extend tests/swarm-orchestrator.test.mjs or add a suite), commit, push, watch CI, relaunch. Task-side problem (ambiguous issue, scope creep) → report to the user, leave the issue In Progress — never fabricate a close.
  3. Known landmines (do not re-learn them the hard way):
    • git push on a branch without upstream silently no-ops — always push origin main explicitly from your own commits;
    • never run the swarm from a non-main branch; checkout -B recreates task branches and the reflog-rescue (v1.2) recovers lost work automatically;
    • the executor runs with opencode run --auto (owner-directed autonomy); all other roles are permission-gated on purpose.

5. Report

When the run ends, summarize to the user in THEIR language: which issues were picked, what shipped (version, PR, CI runs, prod-QA verdicts), what was closed and why. Linear is the source of truth — quote its final states. Then stop; never start another run without the user asking (or a queued --max they requested).

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.