agentleFS
Sign inStart free
southglorysouthglory/system-agents-template/.agents/skills/system-agents-turn/SKILL.md

system-agents-turn

Skill2 starsChanged 5 months agoMITView on GitHub

Use when Codex is acting as a participant in a system-agents turn-based workflow, reading chatrooms, claiming tasks, executing assigned work, or reporting turn results.

The file

Show its rules and text

We split this skill into the rules it gives an agent

13 rules. About 0 are common, 4 are distinctive, and 9 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 9

    Very specific

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

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

    Use this workflow when operating inside a repository that contains chatrooms/, tasks/board.yaml, and the system-agents protocols.

    line 8
  2. Very specific:

    Read chatrooms/PROTOCOL.md and tasks/PROTOCOL.md if they are not already fresh in context.

    line 12
  3. Very specific:

    Read tasks/board.yaml as state only.

    line 14
  4. Distinctive:

    If the phase is ambiguous, summarize the latest chatroom and board state and ask the user which phase to run.

    line 20
  5. Very specific:

    Write messages under chatrooms/{room}/ using this filename format:

    line 24
  6. Very specific:

    Use UTC time. Include frontmatter with at least from, to, time, type, and subject.

    line 30
  7. Very specific:

    After writing a message, update chatrooms/.read-status/{agent}.json so the room points at the message just written.

    line 34
  8. Very specific:

    Check unread messages in general and any relevant 1:1 rooms.

    line 38
  9. Very specific:

    Do not edit tasks/board.yaml.

    line 47
  10. Distinctive:

    Do not edit existing chatroom messages.

    line 48
  11. Distinctive:

    Do not do execution work during planning phase.

    line 49
  12. Very specific:

    Do not send task-claim during execution phase.

    line 50
  13. Distinctive:

    Use hooks as mechanical enforcement, but still follow the protocol intentionally.

    line 51

≈ “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

System Agents Turn

Use this workflow when operating inside a repository that contains chatrooms/, tasks/board.yaml, and the system-agents protocols.

Before Acting

  1. Read chatrooms/PROTOCOL.md and tasks/PROTOCOL.md if they are not already fresh in context.
  2. Determine the current agent name from the current directory, role.md, or explicit user instruction.
  3. Read tasks/board.yaml as state only. Never edit it directly.

Decide The Phase

  • Planning phase: read unread messages, inspect the board, claim work with a task-claim message, and stop before code or document edits.
  • Execution phase: read bot updates, perform assigned work, report progress or completion, then end the turn.
  • If the phase is ambiguous, summarize the latest chatroom and board state and ask the user which phase to run.

Message Writing

Write messages under chatrooms/{room}/ using this filename format:

YYYY-MM-DD_HHMMSS_agent.md

Use UTC time. Include frontmatter with at least from, to, time, type, and subject.

For task-update, task-done, and task-claim, include ref: T-NNN. The ref field is the only task identity the bot uses.

After writing a message, update chatrooms/.read-status/{agent}.json so the room points at the message just written.

Turn Loop

  1. Check unread messages in general and any relevant 1:1 rooms.
  2. Read tasks/board.yaml.
  3. In planning phase, send claims or task creation requests only.
  4. In execution phase, do the assigned work, then send task-update or task-done.
  5. Send a turn-end message to general.
  6. Tell the user what changed and which bot phase should run next.

Safety Rules

  • Do not edit tasks/board.yaml.
  • Do not edit existing chatroom messages.
  • Do not do execution work during planning phase.
  • Do not send task-claim during execution phase.
  • Use hooks as mechanical enforcement, but still follow the protocol intentionally.

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