blog-geo

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted blog content to generate a citation readiness report, creating a surface for indirect prompt injection. \n
  • Ingestion points: The skill reads full content, headings, and metadata from a file path provided by the user, as detailed in the Audit Process (SKILL.md). \n
  • Boundary markers: No explicit delimiters or instructions are used to separate user-provided content from the audit logic or to prevent the agent from following instructions embedded within the analyzed file. \n
  • Capability inventory: The skill is capable of executing local Python scripts and interacting with Google Search Console APIs (via the blog-google skill). \n
  • Sanitization: There is no indication that the processed content is sanitized or filtered before analysis. \n- [DYNAMIC_EXECUTION]: The skill invokes local Python scripts to perform scoring and data retrieval using parameters derived from user input. \n
  • Evidence: The skill calls 'python3 scripts/ai_citation_score.py ' and 'python3 skills/blog-google/scripts/run.py' in the optional sections of SKILL.md. These are standard script executions using a user-provided file path.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 04:57 AM
Security Audit — agent-trust-hub — blog-geo