| name | loopify-governance |
| description | Use when adding budgets, approvals, rollback plans, data/privacy limits, dependency policy, authority boundaries, and harness-change contracts to higher-risk loops. |
loopify-governance
Add governance to higher-risk loops.
When To Use
- A loop can affect production, security, money, data, migrations,
dependencies, public content, or external accounts.
- A loop can mutate its own harness, checks, prompts, tools, or agent
configuration.
- A loop needs budgets, approvals, rollback, or stronger authority boundaries.
Do Not Use
- Do not slow down low-risk loops with unnecessary ceremony.
- Do not let an agent silently widen its own authority.
- Do not treat approval as durable unless it is recorded.
Workflow
- Read the Loop Contract.
- Identify risk class with
references/risk-classes.md.
- Add allowed paths, denied paths, budgets, approvals, rollback, data/privacy
constraints, and dependency policy.
- Use
references/authority-boundaries.md for boundary language.
- Add a harness-change contract if prompts, checks, tools, context builders,
memory, or permissions can change.
- Update stop and escalation rules.
- Add review gates.
Output
Governance addendum, updated Loop Contract, or harness-change contract.
Validation
- High-risk actions require explicit approval.
- Rollback is named where mutation is possible.
- Harness mutations are reviewed as future-behavior changes.
Example
Use loopify-governance to add budgets, denied paths, and a harness-change contract to .loopify/loops/001-papertrail/loop-contract.md.
