Red-Team Eval Authoring for AI Agents: Enhance Security Tools

Red-Team Eval Authoring for AI Agents: Enhance Security Tools

This skill assists developers and teams in creating, reviewing, and managing red-team evaluation artifacts for AI agents using platforms such as Claude Code, Cursor, and Codex. It provides the structured procedures and templates necessary to enhance security assessments by authoring attack templates, grading rubrics, and test metadata.

What this skill does

• Attack Template Authoring: Write structured attack prompts, ensuring one attack per line or machine-parseable record format.
• Grader Rubric Creation: Develop detailed rubrics with structured JSON output to evaluate AI model responses.
• Grader Input Standardization: Ensure that grader inputs are consistent, using predefined fields such as `user_query`, `system_purpose`, `model_output`, and `allowed_entities`.
• Workflow Management: Define harm class details and establish pass/fail boundaries for safety evaluation.

Who it is for

This skill is designed for developers, security engineers, and red-team analysts working with AI coding agents who need to author and manage security evaluations.

Use cases

• Adding new red-team plugins or graders for security analysis.
• Editing attack templates or reviewing evaluation rubrics for accuracy in detecting false negatives/positives.
• Conducting safety evaluations for AI tools in multimodal or tool-use contexts.

Technical details

• Integrates within existing eval frameworks and repositories.
• Supports deterministic shape checks for templates before rubric creation.
• Requires confirmation before running networked evaluations or large red-team suites involving paid model graders.
• Compatibility with AI platforms such as Claude Code, Cursor, and Codex.

For authorised security testing, defensive research, and educational use only.

Source & Licence

This package is built on open-source work published by yeaight7 (yeaight7/agent-powerups) and distributed under Apache-2.0. The original licence text and copyright notice are included in your download.

Personal and commercial use, modification and redistribution are permitted under the Apache License 2.0, which also includes an express patent grant. Attribution and any NOTICE file must be retained.

Your purchase covers curation, licence verification, packaging, documentation and instant delivery. It does not grant exclusive rights to the underlying open-source code, which remains available under its original licence.

Delivery & Support

• Delivery: instant — a secure download link is emailed to you as soon as payment is confirmed.
• Format: ZIP archive containing the skill files, documentation and the original licence.
• Support: support@mcpcart.com — we aim to reply within 2 business days.
• Updates: updates are included only where stated on this page.

Refunds

This is a digital product delivered immediately after purchase. By completing your order you request immediate delivery and acknowledge that, once the download has been accessed, the statutory right to cancel no longer applies to the extent permitted by law. Refund requests are handled in accordance with our published Refund Policy.

Claude, Codex, Gemini and Cursor are trademarks of their respective owners. MCP Cart is an independent marketplace and is not affiliated with, endorsed by, or sponsored by any of them. Compatibility references describe interoperability only.