---
name: governed-agent-execution-kit
description: Create auditable plan/review/approve execution packs for supervised AI agent work in regulated, public-sector, enterprise, or high-risk environments. Use when a task needs clean diffs, approval gates, evidence logs, rollback notes, policy checks, or a structured execution template before an agent changes files, infrastructure, customer data, or operational workflows.
---
Governed Agent Execution Kit
Workflow
1. Define the task, risk level, approver, and affected systems.
2. Run `scripts/governed_execution.py` to create an execution pack.
3. Complete the plan and approval checklist before changes.
4. Capture commands, diffs, test evidence, and rollback notes after execution.
5. Record approval evidence externally; the generated pack is audit evidence, not approval.
Parameters
→ `--task TEXT` required objective.
→ `--risk low|medium|high|critical` governance level.
→ `--approver TEXT` named approval role or person.
→ `--system TEXT` repeatable affected systems.
→ `--output PATH` optional file.
→ `--format json|md` output format.
→ `--self-test` run validation checks.
Output
A plan/review/approval template with risk gates, evidence fields, rollback plan, and completion criteria.
Release evidence
Reviewed in the 2026-06-15 GetAgentIQ warehouse propagation cycle.