agentleFS
apokamoapokamo/kaji/.claude/skills/herdr-kaji-launch/SKILL.md

herdr-kaji-launch

Skill16 starsChanged 2 days agoApache-2.0View on GitHub

Open a sibling Herdr pane and launch kaji there as a real interactive command. Use when a Codex or Claude Code agent running inside Herdr is asked to start, show, or hand off a kaji workflow in another pane; never use Claude Code print mode (`-p`).

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 1 is distinctive, and 6 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.

  • 6

    Very specific

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

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

    Use the kaji command requested by the user.

    line 12
  2. Very specific:

    Check HERDR_ENV=1 and a non-empty HERDR_PANE_ID.

    line 21
  3. Very specific:

    If the release-matched Herdr skill is not already in context, run herdr --skill and read its complete output before any pane operation.

    line 24
  4. Very specific:

    Do not use agent start for kaji.

    line 55
  5. Very specific:

    Do not use terminal output matching as workflow completion authority.

    line 57
  6. Very specific:

    Do not install Herdr integrations or plugins as part of launching kaji.

    line 59
  7. Distinctive:

    Do not close, prune, or reuse any pane whose ID did not come from this split response.

    line 60

≈ “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

Herdr Kaji Launch

Launch kaji in a response-identified sibling pane while preserving the caller's cwd and focus.

Input

Use the kaji command requested by the user. For kaji run, require a workflow path and Issue ID. Unless the user explicitly supplied equivalent options, add:

--agent-runner interactive-terminal --interactive-terminal-backend herdr

Procedure

  1. Check HERDR_ENV=1 and a non-empty HERDR_PANE_ID. If either check fails, stop and tell the user to start this agent inside Herdr. Do not inspect or control the UI-focused session from outside Herdr.
  2. If the release-matched Herdr skill is not already in context, run herdr --skill and read its complete output before any pane operation. Follow its guardrails if they are stricter than this skill.
  3. Resolve the intended cwd. Default to the caller's current working directory; use a different cwd only when the user supplied it or the target worktree was resolved explicitly.
  4. Split from the explicit caller pane with --no-focus and parse the new pane ID from the JSON response. Never predict an ID and never omit the target.

    herdr pane split "$HERDR_PANE_ID" \
      --direction right \
      --ratio 0.5 \
      --cwd "$PWD" \
      --no-focus
  5. Build one shell-quoted interactive kaji command from the user's arguments and run it in the response-derived pane ID.

    herdr pane run "<response-pane-id>" \
      "kaji run <workflow> <issue> --agent-runner interactive-terminal --interactive-terminal-backend herdr"
  6. Report the pane ID, cwd, and exact kaji argv to the user. Leave the pane open for interaction. Close it only if the user asks, and only after re-reading that exact pane ID.

Invariants

  • Launch kaji as a real interactive command. Do not invoke claude -p, Claude Code print mode, or any headless substitute for this path.
  • Do not use agent start for kaji; kaji is the command running in the pane, not the Herdr-recognized coding-agent occupant.
  • Do not use terminal output matching as workflow completion authority. The nested interactive runner advances on its own verdict.yaml artifacts.
  • Do not install Herdr integrations or plugins as part of launching kaji.
  • Do not close, prune, or reuse any pane whose ID did not come from this split response.

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.