# Approval Skill

> Approve or reject after review

- Skill: `cybertheory/approval-skill` (Agent Skill)
- Install (CLI): `npx skillmds@latest add cybertheory/approval-skill`
- Raw SKILL.md: https://api.skillmd.com/api/skills/cybertheory/approval-skill/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: cybertheory (https://skillmd.com/u/cybertheory)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/cybertheory/approval-skill

---


# Approval Skill

Use this skill when the workflow state is **REVIEW** or **FINAL_DECISION**.

## Options

- **approve**: Mark the item as approved. No body required unless the workflow expects optional fields.
- **reject**: Reject with a reason. Body must include:
  - `reason` (string, required): Human-readable reason for rejection.

## Example (reject)

```json
{
  "reason": "Audit found 2 critical issues; requires remediation."
}
```

Always prefer the **approve** action when the audit is clear and no escalation is needed.