mr-description-generator

Pass

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: LOWCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION] (LOW): The skill instructs the agent to execute local shell commands (git log, git diff). These are standard read-only operations used to gather context for the MR description, but they do involve direct interaction with the host's shell.
  • [PROMPT_INJECTION] (LOW): Indirect Prompt Injection Surface Detected.
  • Ingestion points: The skill ingests untrusted data from the output of git diff develop...HEAD in Step 1.
  • Boundary markers: Absent. The instructions do not define delimiters to separate the diff data from the agent's instructions.
  • Capability inventory: The skill possesses command execution capabilities (git commands) and text generation capabilities.
  • Sanitization: Absent. There are no instructions to the agent to ignore or sanitize embedded instructions found within code comments, string literals, or documentation in the diff output.
  • Risk: An attacker could commit code containing "hidden" instructions (e.g., in a comment) that attempt to manipulate the AI's categorization of changes or the content of the final MR description.
Audit Metadata
Risk Level
LOW
Analyzed
Feb 16, 2026, 01:36 PM