Messaging the Human
Lets an agent send clear, safe decision requests or updates to a human through OpenRig's messaging channels.
- Repository
- mvschwarz/openrig
- Stars
- 8,553
- Licence
- Apache-2.0
- Files
- 1 file
What it does
When a project needs a human decision or update, or a reply to a human must be routed back to the right work, this skill teaches the agent how to find the right person or channel, check that delivery is ready, and write a short phone-friendly brief with a recommendation and the choice requested. It also covers urgent alerts, a 'decision room' some people set up so the team can decide on their behalf in pre-approved areas, and recording those decisions honestly. The skill does not decide when a human must be contacted; that comes from the project's own policy.
Great for
- Developers running multi-agent OpenRig projects who need agents to escalate decisions to people
- Teams that want consistent, concise approval requests and status updates sent to Slack or the OpenRig TUI
- Anyone configuring how agents should handle human replies and delivery failures
Skip it if
- Solo use outside OpenRig, since it depends on the rig CLI and a configured gateway (the agent cannot message anyone without that setup)
- Situations where the agent should decide without oversight; the skill warns against converting solvable technical problems into human gates, and mode-based decisions are only allowed when the person's own stated room plainly covers the case
- Sending sensitive data to channels, because the files do not describe how content is protected in transit or at rest
Try asking
- “Ask me to approve an instance update and summarize the tradeoffs before sending it to the human.”
- “Check whether a Slack reply from the human has been delivered and resume the paused work.”
- “Send the human a status update saying the repaired view is ready, with no action needed.”
- “Check whether the human has a decision room set up before asking them about this change.”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 1 original files, the Apache-2.0 licence and an attribution note.
messaging-the-human.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip messaging-the-human.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip messaging-the-human.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Put the full decision in the main brief; the skill says optional detail replies are supplementary, not where essential choices belong.
- Use --alert only for truly urgent asks such as sign-ins or expiring approvals, to avoid interrupting people unnecessarily.
- Check the gateway readiness output before relying on delivery; 'indeterminate' is not treated as ready.
What's inside
- SKILL.md
Redistributed unmodified from mvschwarz/openrig under the Apache-2.0 licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.