embodiment-description

Pass

Audited by Gen Agent Trust Hub on Nov 10, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from project files to generate patent text, creating a surface for indirect prompt injection findings.\n- Ingestion points: Data is read from patent/INVENTION_DISCLOSURE.md, patent/CLAIMS.md, user-provided figures, and patent/figures/numeral_index.md.\n- Boundary markers: Absent. The instructions do not define delimiters or provide warnings to the agent to disregard instructions that might be embedded within the ingested patent documents.\n- Capability inventory: The skill is configured with access to Bash(*), Read, Write, Edit, Grep, and Glob tools. Access to a shell (Bash) provides a significant execution capability if the agent is manipulated by injected content.\n- Sanitization: Absent. No validation or sanitization steps are defined for the content extracted from external files before it is incorporated into the prompt context.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 10, 2026, 09:36 AM