code-simplifier

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill operates by analyzing and refactoring user-provided source code, which constitutes an inherent surface for indirect prompt injection attacks.
  • Ingestion points: Untrusted source code provided for simplification, which is the primary data source for all 45 rules defined in the references directory.
  • Boundary markers: The skill's instructions do not specify the use of delimiters or provide the agent with directives to ignore instructions that may be embedded within the source code it processes.
  • Capability inventory: The AI agent implementing these simplification rules typically has access to the local file system and the ability to execute terminal commands.
  • Sanitization: There are no provisions or instructions within the skill for sanitizing, validating, or escaping the content of the processed code before it is analyzed or modified.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 01:00 PM
Security Audit — agent-trust-hub — code-simplifier