court-auction-notice-search
Pass
Audited by Gen Agent Trust Hub on May 4, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes data fetched from an external government website (
courtauction.go.kr), which constitutes an indirect prompt injection surface. While the source is a public legal database, the skill lacks explicit instructions for the agent to ignore or sanitize embedded commands within the fetched content. - Ingestion points: External data is ingested via POST requests to various endpoints on
courtauction.go.kr(e.g.,selectRletDspslPbanc.on). - Boundary markers: None detected in the instructions; the agent is not explicitly told to delimit or ignore instructions within the scraped JSON data.
- Capability inventory: The skill is read-only and restricted to network GET/POST requests and JSON processing. No file-write or shell execution capabilities are exposed.
- Sanitization: The skill description does not mention sanitizing or escaping the retrieved data before presenting it to the agent.
Audit Metadata