wayne930242wayne930242/straw-boss/skills/notifying-main-agent/SKILL.md

notifying-main-agent

Skill5 starsChanged 54 days agoMITView on GitHub

Use when a dispatched agent routes a question, progress, or status to its main agent.

In this repository18 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- agent-feedbackSkill
- asking-peer-agentsSkill
- boss-assistantSkill
- boss-saySkill
- bringing-coworkerSkill
- choosing-graphSkill
- contacting-orchestratorsSkill
- create-great-harnessSkill
- dispatching-workSkill
- handoff-orchestratorSkill
- i-am-orchestratorSkill
- initSkill
- peeking-workSkill
- reporting-to-userSkill
- resolving-appSkill
- shipping-taskSkill

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 1 is distinctive, and 3 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.
3
Very specific
Tied to their project's names, files or tools. Probably skip.

Common
Distinctive
Very specific

Distinctive:
Ask it only for integrated context or a coordinator-owned action result.
line 9
Very specific:
Put detailed context or evidence in repeatable --ref.
line 13
Very specific:
If the coordinator's answer becomes blocking, report awaiting-main-agent.
line 29
Very specific:
If no live endpoint exists, record the same detail with report-progress.py.
line 57

≈ “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

Overview

Discuss work details and authorization directly with the user; the main agent accepts those decisions. Ask it only for integrated context or a coordinator-owned action result. Address every operation by your instruction path.

Live bodies carry one delta in at most two sentences. Put detailed context or evidence in repeatable --ref; transport adds identity, intent, and correlation.

Ask the main agent

From awaiting-user-input or awaiting-authorization, report awaiting-main-agent with the needed action or fact before asking; transport enforces this order, while progress notes preserve the checkpoint.

uv run --script "${CLAUDE_PLUGIN_ROOT}/scripts/send-dispatch-message.py" \
  --instruction-path <your exact instruction path> \
  --to main --intent question \
  --message "<one integration question; include expected reply shape>" \
  --ref "<source or artifact when needed>"

Continue independent work. If the coordinator's answer becomes blocking, report awaiting-main-agent; a user-owned question remains awaiting-user-input. awaiting-main-agent means a coordinator action or fact is required. Waiting on your own subagent, review, or CI is progress, recorded as a wait under Report status.

Report status

At every checkpoint and terminal outcome, make exactly one call:

uv run --script "${CLAUDE_PLUGIN_ROOT}/scripts/report-task-status.py" \
  --instruction-path <your exact instruction path> \
  --status <done|failed|awaiting-authorization|awaiting-user-input|awaiting-main-agent> \
  --note "<outcome or exact unblock>" --ref "<proof when needed>"

The command persists status before notifying the recorded main-agent Herdr endpoint. Delivery failure is surfaced and leaves durable state for watcher recovery. A reply to an awaiting-main-agent checkpoint leaves status unchanged and only marks that checkpoint resolved; before stopping, report a fresh status here or record a wait, since the resolved one no longer counts. To end a turn while your own CI watch, subagent, review, or scheduled wake will resume you, run report-progress.py with --note "<current state>" --waiting-on "<what resumes you>"; each wait note covers one stop.

Report a feature-branch push

uv run --script "${CLAUDE_PLUGIN_ROOT}/scripts/send-dispatch-message.py" \
  --instruction-path <your exact instruction path> \
  --to main --intent inform \
  --message "Pushed <branch>; continuing." --ref "<MR/PR>"

If no live endpoint exists, record the same detail with report-progress.py. Continue immediately.

Complete when: the intended message was delivered, or terminal status was persisted and its Herdr notification succeeded or visibly failed.

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