scanning-input-validation-practices

Pass

Audited by Gen Agent Trust Hub on Mar 14, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is inherently vulnerable to indirect prompt injection because it processes untrusted external codebase data which may contain adversarial instructions designed to influence agent behavior. \n
  • Ingestion points: Recursive directory scanning and file reading are performed by the scripts 'codeql_scan.sh' and 'semgrep_scan.sh'. \n
  • Boundary markers: No clear boundaries or instructions are provided to the agent to distinguish between the code being analyzed and potential malicious instructions embedded within it. \n
  • Capability inventory: The skill utilizes powerful tools including 'Bash(security:*)', 'Read', 'Write', and 'Edit', which could be leveraged if an injection is successful. \n
  • Sanitization: There is no evidence of content sanitization or instruction filtering applied to the scanned files before they are read into the agent's context.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 14, 2026, 03:14 AM