require-external-communication-approval
Skill1 starsChanged 2 months agoMITView on GitHub
Check authorization before sending a direct message, review, comment, or other human-facing communication.
In this repository16 other files it gives its agents.
- adversarial-refinement-loopSkill
- atomic-commitsSkill
- audio-verificationSkill
- build-without-fearSkill
- comfyui-assetsSkill
- diagnosing-bugsSkill
- finish-the-listSkill
- frontend-designSkill
- headless-terminal-workflowSkill
- parallelize-independent-workSkill
- prefer-second-monitorSkill
- state-of-the-art-firstSkill
- try-it-firstSkill
- visual-verificationSkill
- web-design-guidelinesSkill
- writing-for-agentsSkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 4 are distinctive, and 0 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.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 10
Never directly send, post, or reply to another person without the user's explicit approval for that specific communication.
- Distinctive: line 24
When otherwise authorized by the user's task, proceed without a separate communication confirmation for:
- Distinctive: line 36
Before asking, check whether the current request or prior session instructions already explicitly authorize this exact communication.
- Distinctive: line 46
Ask for explicit approval.
≈ “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.
Files
Require External Communication Approval
Hard rule
Never directly send, post, or reply to another person without the user's explicit approval for that specific communication.
Require approval before:
- Posting or replying to a pull-request, review, issue, or discussion comment.
- Sending Slack, Teams, email, chat, direct messages, or similar messages.
- Submitting reviews, reactions, mentions, assignments, or reviewer requests.
- Resolving or dismissing another person's review or discussion.
Drafting text without sending it is allowed.
Do not block operational work
When otherwise authorized by the user's task, proceed without a separate communication confirmation for:
- Commits, pushes, merges, branch operations, and repository settings.
- Workflow triggers, reruns, cancellations, builds, tests, and deployments.
- Releases, infrastructure changes, and automatic bot-generated pull requests.
- Creating or updating pull requests that the user asked Codex to manage, provided Codex does not add reviewers, mentions, or follow-up comments.
Automatic service notifications caused by these operational actions do not count as Codex directly messaging another person.
Before asking, check whether the current request or prior session instructions already explicitly authorize this exact communication. Existing authorization persists; ask only for missing approval or a material change to recipients, destination, or content. Immediately before sending, verify the planned action still matches that authorization.
Approval workflow when authorization is missing
- Draft the proposed communication without sending it.
- Show the user the destination, recipients, and exact text or action.
- Ask for explicit approval.
- Send only the approved communication without adding wording or recipients.
Requests to fix code, manage a pull request, or perform a deployment authorize the operational work, but do not authorize a separate human-facing reply.
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
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.