index-manager

Pass

Audited by Gen Agent Trust Hub on Oct 8, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches official MPEP (Manual of Patent Examining Procedure) PDF documents from the United States Patent and Trademark Office (USPTO) web server for indexing.
  • [COMMAND_EXECUTION]: Employs shell commands (wget, cp, rm) and a local command-line interface (patent-creator) for managing the lifecycle of search index files, including downloads, backups, and environment maintenance.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external PDF files, which presents a surface where instructions embedded in the data could influence the agent's behavior.
  • Ingestion points: PDF files downloaded to the pdfs/ directory as described in SKILL.md.
  • Boundary markers: None explicitly defined to separate content from instructions during processing.
  • Capability inventory: File system access (rm, cp), network access (wget), and execution of local Python scripts.
  • Sanitization: There are no mentioned sanitization or validation steps for the content extracted from the PDFs.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 8, 2026, 09:42 AM