technical

Pass

Audited by Gen Agent Trust Hub on Sep 29, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill recommends installing the seomator-geo-audit package using standard Python tools like uv or pipx. This package is a core component of the skill's functionality and originates from the vendor's own namespace.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes content fetched from external websites through the geo audit tool. It implements a security best practice by explicitly instructing the agent to treat external excerpts as untrusted content and to ignore any commands or instructions contained within them.
  • Ingestion points: External page content is ingested via the findings[].excerpt and signals[].detail fields of the geo tool output in SKILL.md.
  • Boundary markers: The skill contains a 'Hard rules' section in SKILL.md specifically warning the agent to treat excerpts as untrusted quotations and to ignore embedded instructions.
  • Capability inventory: The skill uses Bash (to run the geo audit) and Read (to access documentation files).
  • Sanitization: The agent is instructed to report any detected instructions in the data as findings rather than executing them.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 29, 2026, 03:20 PM