file-intel

Warn

Audited by Gen Agent Trust Hub on Mar 29, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to run a local Python script scripts/process_files_with_gemini.py via the bash shell. Because the script's source code is not included in the skill definition, its security posture and actions on the local filesystem cannot be verified.
  • [COMMAND_EXECUTION]: The skill uses the open command to automatically open output directories. While used here for user convenience, the open command can be used to execute arbitrary files if paths are manipulated.
  • [PROMPT_INJECTION]: The skill is designed to process and summarize content from untrusted external files (PDF, DOCX, XLSX, etc.). This creates a vulnerability to indirect prompt injection, where an attacker could embed malicious instructions within these files to hijack the agent's behavior during the processing phase.
  • Ingestion points: Local folders and files processed by the script (SKILL.md).
  • Boundary markers: No delimiters or ignore instructions are specified for the processed content (SKILL.md).
  • Capability inventory: The agent has shell execution capabilities (python, open) (SKILL.md).
  • Sanitization: No evidence of input validation or sanitization is provided (SKILL.md).
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 29, 2026, 01:27 AM