munich-apartment-search
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [DATA_EXFILTRATION]: The skill instructs the agent to access sensitive information stored in the
self-sourcedirectory, specifically including the user's current home location and stable identity or contact facts. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and evaluate untrusted data from external sources, creating a surface for indirect prompt injection.
- Ingestion points: The agent is directed to retrieve and manually filter titles and descriptions from external apartment listing platforms such as ImmoScout24 (found in
SKILL.md). - Boundary markers: The instructions do not define delimiters or provide specific warnings for the agent to ignore instructions that might be hidden within listing descriptions.
- Capability inventory: The skill uses the agent's inherent capabilities to browse the web and process text, which could be exploited if a listing contains malicious instructions.
- Sanitization: There are no procedures defined for validating or sanitizing the content of external listings before the agent processes them.
Audit Metadata