skills/lev-os/agents/geo-schema/Gen Agent Trust Hub

geo-schema

Pass

Audited by Gen Agent Trust Hub on Mar 21, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill employs curl or WebFetch to retrieve the HTML content of target websites for structured data auditing. This network operation is essential for the skill's stated purpose.
  • [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it ingests and processes untrusted HTML content from the web. The security risk is minimal because the skill is designed for passive validation and reporting rather than executing retrieved content.
  • Ingestion points: Target website HTML fetched during Step 1 (SKILL.md).
  • Boundary markers: None present to delimit untrusted web content from instructions.
  • Capability inventory: Network read access via curl. No file writing, privilege escalation, or shell execution capabilities identified.
  • Sanitization: Performs syntactic validation of JSON-LD data structures but does not explicitly sanitize text content for embedded instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 21, 2026, 06:48 PM