test-fixing

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses local shell commands including make test, uv run pytest, and git diff to identify failures, verify fixes, and review recent changes. These are standard operations for development and testing workflows.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data, creating an attack surface for indirect prompt injection.
  • Ingestion points: The agent reads and analyzes the output of make test, pytest, and git diff (SKILL.md).
  • Boundary markers: There are no explicit delimiters or specific instructions to ignore malicious content embedded within test failure messages or code diffs.
  • Capability inventory: The skill possesses the ability to execute shell commands (make, pytest, git) and perform file system writes via the agent's edit tool.
  • Sanitization: The skill does not implement sanitization or validation of the test output data before it is processed by the AI agent's reasoning engine.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 07:19 AM
Security Audit — agent-trust-hub — test-fixing