code-reviewer

Pass

Audited by Gen Agent Trust Hub on Sep 6, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests user-supplied code as a primary input, creating a potential surface for indirect prompt injection where instructions hidden in the code could influence agent behavior.
  • Ingestion points: User-provided code via the variable in SKILL.md.
  • Boundary markers: Employs tags to separate untrusted content from the prompt instructions.
  • Capability inventory: The skill operates with the agent's default permissions, which often include shell and file system access.
  • Sanitization: No sanitization or filtering is applied to the code before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 6, 2026, 08:56 AM
Security Audit — agent-trust-hub — code-reviewer