agentleFS
vosslabvosslab/vosslab-skills/skills/management/stay-busy/SKILL.md

stay-busy

Skill5 starsChanged 24 days agoMITView on GitHub

Keep an agent manager productive through safe, evidence-producing follow-on work. Use for `/stay-busy`, idle orchestrators, waiting subagents, or active delegated workflows. Generates bounded workstreams and fallbacks; not productivity advice.

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. Very specific:

    Use this only inside an active delegate-manager-to-subagents workflow when the obvious execution queue is genuinely empty and a safe, …

    line 8
  2. Distinctive:

    Every selected task must produce evidence or remove a blocker.

    line 21
  3. Distinctive:

    Read references/operating_modes.md for away-mode decisions, safe defaults, and situation routing.

    line 35
  4. Distinctive:

    Read references/scaling_and_cleanup.md to select the tier, apply an away-mode bump, and close stale excess work.

    line 37
  5. Distinctive:

    Read references/ideas_checklist.md to choose evidence-producing work and reject busywork.

    line 39
  6. Distinctive:

    Create one task per chosen workstream, including its goal, owner boundary, done-when evidence, canonical status requirement, artifact path, …

    line 54

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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

Stay busy

Use this only inside an active delegate-manager-to-subagents workflow when the obvious execution queue is genuinely empty and a safe, project-relevant workstream remains. It generates workstream-shaped tasks; the manager owns dispatch and completion.

Non-negotiable boundary

Finish the obvious before expanding. Inspect every active or paused plan task first. If any task needs implementation, verification, documentation, or handoff, queue and close that task before creating a new workstream. Do not use this skill to avoid final verification, to inflate scope, or to turn a targeted request into speculative production changes.

Every selected task must produce evidence or remove a blocker. It carries:

  • one canonical status: DONE, DONE_WITH_CONCERNS, NEEDS_CONTEXT, or BLOCKED; and
  • one inspectable artifact path, such as a report, screenshot, command log, JSON record, or before/after metric.

SIDE QUEST is an optional task annotation, never a status. It remains project-related and distinct from production-ready work.

Decide before dispatch

  1. Confirm the active approved plan, current idle state, and the clean finish-before-expanding gate.
  2. Read references/operating_modes.md for away-mode decisions, safe defaults, and situation routing.
  3. Read references/scaling_and_cleanup.md to select the tier, apply an away-mode bump, and close stale excess work.
  4. Read references/ideas_checklist.md to choose evidence-producing work and reject busywork. Then load only the selected workstream's requirements through references/workstream_templates.md.
  5. Attach an explicit blocked fallback and use references/boundaries.md for every authority or ask-only decision.

Away mode means the manager documents reversible assumptions and continues, then produces finished artifacts rather than questions for the absent user. A real boundary remains a boundary; route it according to the references instead of inventing authority.

Emit a usable handoff

Create one task per chosen workstream, including its goal, owner boundary, done-when evidence, canonical status requirement, artifact path, and blocked fallback. Use the standard output template from references/workstream_templates.md, with the allowed-without-asking and ask-only lists from references/boundaries.md. Return the task list to delegate-manager-to-subagents for dispatch.

For lifecycle rationale, overnight examples, and sibling-skill composition, read references/big_picture.md. The routed references intentionally hold conditional detail so this entrypoint remains a small control plane.

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.