saboteur-adversarial-review

Identify security vulnerabilities in pull requests using adversarial review patterns.

3|2|Updated May 12, 2026
One-click install
npx skills add https://github.com/kriscendobot/garden --skill saboteur-adversarial-review
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: saboteur-adversarial-review
Source: https://github.com/kriscendobot/garden/tree/main/skills/saboteur-adversarial-review
Command: npx skills add https://github.com/kriscendobot/garden --skill saboteur-adversarial-review

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill addresses the blind spots in standard code reviews by simulating an adversarial perspective, identifying security vulnerabilities that typical type-checkers and unit tests miss.

Core Features & Use Cases

  • Adversarial Pattern Matching: Automatically checks for common attack vectors like symlink-based path traversal, delimiter injection, and fail-open error handling.
  • Automated Security Gate: Integrates directly into the gardening panel to flag high-risk code patterns before they are merged.
  • Use Case: When a developer submits a PR involving filesystem path manipulation or environment variable derivation, this skill proactively identifies potential sandbox escapes or injection vulnerabilities.

Quick Start

Run the saboteur adversarial review on the current pull request to identify potential security vulnerabilities in the proposed changes.

Frequently Asked Questions about saboteur-adversarial-review

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I identify sandbox escape vulnerabilities in pull requests?

Sandbox escape vulnerabilities in pull requests are identified by applying adversarial review patterns to filesystem path manipulation, environment variable derivation, and error handling logic. This detects symlink-based path traversal and fail-open conditions before merging.

What security flaws does adversarial code review catch that static analysis misses?

Adversarial code review catches security flaws that standard static analysis misses by simulating attack vectors like delimiter injection and fail-open error handling. It proactively targets potential sandbox escapes and injection vulnerabilities in proposed code changes.

How do I prevent fail-open conditions and path traversal in my code?

Fail-open conditions and path traversal are prevented by running automated adversarial security checks on pull requests. This deterministic review targets filesystem path manipulation and environment variable derivation logic to flag high-risk code patterns.

Can I automate security checks for symlink attacks in my pull request workflow?

Security checks for symlink-based path traversal attacks can be automated directly within your pull request workflow. The review integrates into the gardening panel to flag high-risk code patterns before they are merged.

Does adversarial security review require integration with the gardening panel?

Adversarial security review requires integration with the gardening panel to execute deterministic checks and provide semantic security feedback. This integration enables automated flagging of high-risk code patterns in pull requests.