invoice-organizer

Pass

Audited by Gen Agent Trust Hub on Sep 23, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructions involve the use of standard shell commands such as find, mkdir, and cp to search for, categorize, and copy files. These operations are limited to the local filesystem and serve the stated objective of document organization.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted data from external invoices and receipts, creating a surface for indirect instructions to be ingested by the agent.
  • Ingestion points: Information extraction from PDF, JPG, and PNG files, including vendor names, descriptions, and metadata.
  • Boundary markers: The skill lacks explicit delimiters or instructions for the agent to ignore potentially malicious content embedded within the files being processed.
  • Capability inventory: The agent is empowered to perform file system operations such as finding files, creating directory structures, and copying files based on extracted information.
  • Sanitization: There is no mention of sanitizing or validating extracted text before using it in shell commands (filenames) or CSV reports, which could lead to command confusion or data corruption if maliciously crafted files are encountered.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 23, 2026, 06:08 PM