Duty loop reviewer
Scene-specific AI workers for Codex and Claude.
npx -y skills add yeyitech/duty-loops --skill duty-loop-reviewerAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Review a Duty Loop's recent runs for usefulness, notification noise, stale context, missing gates, and over-engineering. Use when Codex should inspect BLUEPRINT.md, state, and logs to suggest small improvements to a long-lived AI worker without turning it into a heavy workflow system.
SKILL.md
1.2 KB, 213 tokens by cl100k_base, as published. Nobody here has run it
Duty Loop Reviewer
Review whether a Duty Loop is still useful, quiet, and safe.
Workflow
-
Read
blueprints/<loop-id>/BLUEPRINT.md. -
Run:
npm run review -- <loop-id> npm run status -- <loop-id> -
Inspect recent
quiet,notify,ask, andactdecisions. -
Look for:
- repeated noisy notifications;
- quiet runs that should have noticed material change;
- stale or missing context sources;
- human gates that are too broad or missing;
- business judgment hard-coded into brittle rules;
- scripts that should stay scripts versus model judgment that should stay flexible.
-
Suggest the smallest useful change.
Output
Return:
- current loop health;
- one or two high-impact improvements;
- whether the issue is context, policy, connector, notification, or gate design;
- a concrete patch plan only if the user asks you to edit files.
What ships with it: 1 file
218 B alongside SKILL.md
agents/
- openai.yaml218 B