refine-tests

Pass

Audited by Gen Agent Trust Hub on Sep 24, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes the php artisan test command to run refined tests. This is a standard operation for local testing and does not involve elevated privileges or remote access.
  • [INDIRECT_PROMPT_INJECTION]: The skill reads and analyzes test files that may have been generated or modified. This creates an ingestion surface for untrusted data if the content of those files is controlled by an external actor.
  • Ingestion points: Test files identified through git diff or recent file system changes.
  • Boundary markers: None present.
  • Capability inventory: Shell execution of testing commands.
  • Sanitization: None performed on the file content before analysis.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 24, 2026, 07:33 AM
Security Audit — agent-trust-hub — refine-tests