nyc-landmarks
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill fetches property and landmark data from the NYC Open Data (Socrata) API at data.cityofnewyork.us. This external data is ingested into the agent's context and represents a potential vector for instructions hidden within dataset fields to influence agent behavior.
- Ingestion points: API responses from the LPC Individual Landmark database and PLUTO datasets retrieved via WebFetch in SKILL.md.
- Boundary markers: Absent; the instructions do not specify the use of delimiters or 'ignore embedded instructions' markers when interpolating data from the Socrata API into the final report.
- Capability inventory: The skill utilizes WebFetch for network operations, Write for report generation, Read for internal instruction references, and Bash for general command execution.
- Sanitization: There is no mention of validation or sanitization of the string values returned from the city databases (e.g., lpc_name, lpc_sitede) before they are processed by the agent.
Audit Metadata