naver-land-scouter

Pass

Audited by Gen Agent Trust Hub on Mar 15, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by processing external data from real estate listings.
  • Ingestion points: External data is ingested via API calls in scripts/naver_land_repository.py and processed in various services.
  • Boundary markers: While the output uses Markdown code blocks and <details> tags for structure, it lacks explicit instructions to ignore natural language commands embedded in the retrieved text.
  • Capability inventory: The skill is restricted to making network requests and formatting reports; no execution or file-system modification capabilities were found.
  • Sanitization: The skill uses Pydantic for structural validation, but does not perform content-based sanitization for potential injection attacks in text fields.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 15, 2026, 02:01 AM