patent-novelty-check

Pass

Audited by Gen Agent Trust Hub on May 13, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection through its integration with external data and cross-model tools.
  • Ingestion points: The skill ingests untrusted data from $ARGUMENTS (invention description) and the file patent/PRIOR_ART_REPORT.md.
  • Boundary markers: Absent. In Step 4, the skill interpolates this data into a prompt for mcp__codex__codex without using delimiters or instructions to the model to ignore embedded commands.
  • Capability inventory: The skill possesses extensive capabilities, including Bash(*), WebSearch, WebFetch, and filesystem write access (Write, Edit).
  • Sanitization: No sanitization or validation is performed on the ingested content before it is processed by the external reviewer model.
Audit Metadata
Risk Level
SAFE
Analyzed
May 13, 2026, 10:11 AM