patent

Pass

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [SAFE]: No malicious patterns or security vulnerabilities were identified in the skill's code or instructions. The skill utilizes official and well-known patent search services.- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external patent databases, creating a potential attack surface for indirect prompt injection. However, this is inherent to the skill's primary purpose of patent research.
  • Ingestion points: Data returned from search tools targeting Google Patents, Espacenet, USPTO, and Lens.org in Phase 3.
  • Boundary markers: Absent; the prompt does not specify delimiters or instructions to ignore embedded content within patent data.
  • Capability inventory: Writing to the local file system (DOCX generation) and executing bundled Python scripts.
  • Sanitization: Absent; the skill does not explicitly mention sanitizing or escaping the content fetched from patent repositories.- [COMMAND_EXECUTION]: The skill executes local Python scripts (sub_use_case_router.py, family_resolver.py, citation_tracker.py) to process patent data and track sessions. These scripts are included in the skill package and use only standard library modules to perform their intended data processing functions.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 16, 2026, 01:16 AM
Security Audit — agent-trust-hub — patent