Skip to content
Back to skills

Gatekeeper

ASecurity

Implement a Creator-Reviewer quality gate loop. The Creator generates an artifact, and the Gatekeeper/Reviewer verifies it against quality criteria. If approved, it passes; if rejected, the Creator must perform rework and resubmit. Useful for strict code reviews, verification checks, or quality assurance. Triggers: "게이트키퍼", "리뷰 루프", "검증 게이트", "품질 검증", "gatekeeper", "reviewer loop", "quality gate".

  • 12 stars
  • 0 votes
  • 0 copies
  • 10 views
  • Added September 28, 2026
ai-agentsgo

Works with

  • mcp

Security analysis

A100/100

Scanned September 28, 2026

npx -y skills add fritzprix/libr-agent --skill gatekeeper --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Gatekeeper?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Gatekeeper
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/fritzprix-gatekeeper/badge)](https://www.skillsdirectory.com/skills/fritzprix-gatekeeper)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

SKILL.md
---
name: gatekeeper
description: >
  Implement a Creator-Reviewer quality gate loop. The Creator generates an artifact,
  and the Gatekeeper/Reviewer verifies it against quality criteria. If approved,
  it passes; if rejected, the Creator must perform rework and resubmit.
  Useful for strict code reviews, verification checks, or quality assurance.
  Triggers: "게이트키퍼", "리뷰 루프", "검증 게이트", "품질 검증", "gatekeeper", "reviewer loop", "quality gate".
---

# Gatekeeper

The Gatekeeper pattern sets up a loop between a Creator session and a Reviewer/Gatekeeper session. The reviewer audits the output against explicit quality criteria, approving it for delivery or rejecting it for rework.

## 4-Stage Workflow

```
 ┌───────────┐      Submit Artifact      ┌──────────────┐
 │  Creator  ├──────────────────────────►│  Reviewer/   │
 └─────▲─────┘                           │  Gatekeeper  │
       │                                 └──────┬───────┘
       │            Reject (Rework)             │
       └────────────────────────────────────────┤ Approve ──► Ship
                                                ▼
```

1. **Specs & Criteria**: Setup target task goals for the Creator and quality criteria for the Reviewer. See [criteria.md](references/criteria.md).
2. **Draft & Submit**: The Creator implements the code/artifact in the shared workspace and notifies the Reviewer.
3. **Audit**:
   - The Reviewer evaluates the artifact against criteria.
   - **Approve**: If all criteria pass, the loop exits.
   - **Reject**: List the failed rules and provide actionable feedback.
4. **Rework Loop**: The Creator modifies the files based on the feedback and resubmits. This loops until approved or loop count limit is reached.

## 🛠️ MCP Tools Guide

- **Feedback Delivery**: Inject the reviewer's audit report using `agent__messageToSession` to wake and direct the Creator to rework.
- **Infinite Loop Avoidance**: The parent session must track and enforce a **maximum rework limit (default: 3)** to prevent endless loops over minor stylistic differences. See [rework-limit-examples.md](references/rework-limit-examples.md).

## References

- [Quality criteria](references/criteria.md)
- [Rework limit examples](references/rework-limit-examples.md)
- [Troubleshooting](references/troubleshooting.md)

Files in this skill

  • SKILL.md2.6 KB
  • references/criteria.md1.5 KB
  • references/rework-limit-examples.md1.1 KB
  • references/troubleshooting.md1.4 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

View all in ai-agents →

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.