file-organization

Pass

Audited by Gen Agent Trust Hub on Mar 9, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill provides automated file organization capabilities by reading existing directory structures and file contents. This creates a surface for indirect prompt injection, as an attacker could place a malicious file within a project to influence the agent's behavior during a refactoring task.\n
  • Ingestion points: Project directories and files read via the Read and Glob tools.\n
  • Boundary markers: Instructions do not include boundary markers to isolate processed file content or directives to ignore instructions found within files.\n
  • Capability inventory: The skill allows the use of Bash for command execution and the Write for creating or modifying files.\n
  • Sanitization: No validation or sanitization of external data is specified.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 9, 2026, 11:59 AM