| name | format-verdict |
| description | Use when producing a review verdict that needs clear evidence, risk level, and a compact decision-ready result. |
Verdict Formatting
Auditor outputs a verdict, not an implementation plan.
Required Enums
- scope:
on-demand | risk-triggered | scheduled
- verdict:
approve | approve-with-notes | revise | reject
Required Shape
Always produce:
## Audit Verdict
**Subject:** <what was reviewed>
**Scope:** <on-demand | risk-triggered | scheduled>
**Verdict:** <approve | approve-with-notes | revise | reject>
### Critical Findings
1. <issue>
### Observations
- <non-blocking note>
### Improvement Opportunities
- <concrete workflow or quality improvement>
### Confidence
<high | medium | low> - <brief justification>
### Recommended Action
<what should happen next and who owns it>
### Escalation Needed
<yes | no> - <why>
Rules
- lead with the decision
- make findings evidence-backed
- separate blocking findings from non-blocking observations
- keep the output compact once the verdict is clear
When returning the result through main, prefer:
## Audit Complete
**Subject:** <brief restatement>
**Verdict:** <approve | approve-with-notes | revise | reject>
### For the user
<one-paragraph summary of findings>
### Deliverables
- Nextcloud: <paths to audit reports stored>
- Qdrant: <memories stored, if any>
### Follow-up needed
<what needs to happen next and owner>
Escalate
- if the review request is missing the artifacts needed to support a defensible verdict
