agentleFS
joaoajmatosjoaoajmatos/ze/.claude/skills/ze-continue/SKILL.md

ze-continue

Skill2 starsChanged 47 days agoUnlicenseView on GitHub

Continues Ze development after a pause. It restarts Docker, the Ground servers and the routines, then relaunches each floor's agents on their own sessions, one floor at a time, and sends them back to their pause handoffs. Use when the user asks to continue, resume or pick up where the team left off.

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. Very specific:

    Read before acting: Read the newest journal/pause/<stamp>.md, including its Maestro summary, the matching project_ze_pause_* memory, and …

    line 9
  2. Very specific:

    If forks fail, use exec <command> for single commands and clean up first, as described in the process-hygiene memory.

    line 16
  3. Very specific:

    Infrastructure: Run python3 scripts/process/pausectl.py resume --infra-only.

    line 18
  4. Very specific:

    Run python3 scripts/process/pausectl.py resume --floor Ground, then --floor "<Floor>" for each floor in the summary's order.

    line 27
  5. Very specific:

    After each floor, maestri check every agent.

    line 28
  6. Distinctive:

    Update the score floor states, add a "Resumed <date>" line under the roadmap's pause line, and update your state file.

    line 36

≈ “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

Continue after a pause (Maestro only, from Ground)

  1. Read before acting. Read the newest journal/pause/<stamp>.md, including its Maestro summary, the matching project_ze_pause_* memory, and the PAUSE entries in "Ground Handoffs". Note halted floors: they stay asleep.
  2. Check the machine.

    • Load and the process count: ps -ax | wc -l should be well under the per-user limit of 2666.
    • Free disk space.
    • Kill leaked docker clients: pkill -f "stats --all --no-trunc".

    If forks fail, use exec <command> for single commands and clean up first, as described in the process-hygiene memory.

  3. Infrastructure. Run python3 scripts/process/pausectl.py resume --infra-only. It starts Docker Desktop and waits for it, starts the compose services (Postgres when the pause could not read them), restarts the recorded Ground servers with their env (logs in ~/.maestri-ze/logs/), and re-enables the routines that were on. Verify each one:

    • docker compose exec -T postgres pg_isready -U ze
    • the API answers on its port (/api/v0/version)
    • the web app answers on 5183
    • the Score portal answers on 4173

    Fix anything that fails before you wake agents.

  4. Agents, one floor at a time.
    1. Run python3 scripts/process/pausectl.py resume --floor Ground, then --floor "<Floor>" for each floor in the summary's order. Never name a halted floor. The script retypes each agent's recorded command (claude --resume <session> on its recorded model, with ZE_ROLE), and waits while the load is above --max-load (default 60).
    2. After each floor, maestri check every agent. A Claude TUI means it came back. A bare shell or a "No conversation found" error means you relaunch it by hand: type the printed command, or use --continue in its directory.
    3. Brief every agent that came back with one batch:

      RESUMED after pause <stamp>. Read your PAUSE entry in your floor's Handoffs note and your state file, run git status, then continue from your next step. Reply with one line: resumed: <next step>.

      Floor leads wake their own lanes only when they need them.

  5. Close the loop.
    • Update the score floor states, add a "Resumed <date>" line under the roadmap's pause line, and update your state file.
    • Set "resumed_at" in the snapshot JSON.
    • Report to the user: a table with one row per floor showing where it is and what happens next, the agents that needed a manual relaunch, and anything still broken.

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.