code-testing-agent

Pass

Audited by Gen Agent Trust Hub on Sep 25, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTION
Full Analysis
  • Structured Multi-Agent Pipeline: The skill coordinates several specialized agents (Researcher, Planner, Implementer) to handle the test generation lifecycle, ensuring a thorough analysis of the codebase before implementation.
  • Workspace State Management: It uses a non-stageable directory for intermediate artifacts like research and plans, which helps maintain project cleanliness by keeping internal working data out of version control.
  • Indirect Prompt Injection Surface: The skill analyzes source code in the user's workspace to identify test cases. This ingestion point could potentially be used to influence the agent's behavior if instructions are embedded within the source code or comments.
  • Ingestion points: Source code files and workspace directory structure (SKILL.md).
  • Boundary markers: The skill relies on natural language parsing without explicit delimiters for untrusted code content.
  • Capability inventory: The skill has the ability to write files to the workspace and execute shell commands via test runners.
  • Sanitization: There is no explicit sanitization or filtering of instructions that might be contained within the processed source code.
  • Dynamic Execution of Generated Tests: The skill generates test files and subsequently executes them using project-specific test runners. While this is the intended purpose, it involves the execution of newly created code within the development environment.
  • Command Execution within Development Scope: The skill invokes standard development tools such as git rev-parse and language-specific test commands. These operations are performed to manage state and verify that generated tests are functional and pass correctly.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 25, 2026, 08:48 PM
Security Audit — agent-trust-hub — code-testing-agent