gs-prune-tests

Pass

Audited by Gen Agent Trust Hub on Nov 19, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
  • [COMMAND_EXECUTION]: The skill guides the agent to perform necessary shell operations within the local environment to manage the test suite.
  • Executes git commands (e.g., git symbolic-ref, git remote show, git log) to determine the project's branch structure and identify relevant test files.
  • Runs the project's existing test suite to ensure a functional baseline and verify that changes do not introduce regressions.
  • [INDIRECT_PROMPT_INJECTION]: The skill analyzes user-provided source code and test files to make logic decisions, which constitutes an ingestion point for potentially untrusted data that could influence the agent's behavior.
  • Ingestion points: Test files and production source code in the user's local repository (referenced in SKILL.md).
  • Boundary markers: The instructions do not specify the use of delimiters or 'ignore' instructions when reading the project files.
  • Capability inventory: Includes shell command execution, file system modifications (deletion of code and tests), and running local test scripts.
  • Sanitization: No explicit sanitization or validation of the code content being analyzed is described.
  • [NO_CODE]: The skill consists entirely of natural language instructions and documentation without including any executable script files (e.g., .js, .py, .sh).
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 19, 2026, 04:36 PM