de-residency-advisor
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's core functionality relies on fetching and reading external web pages, which introduces an attack surface for indirect prompt injection where malicious instructions could be embedded in the retrieved content.
- Ingestion points: As described in
references/research-playbook.mdandreferences/sources.md, the skill triggers the host agent to search and fetch content from various external URLs and German authority portals. - Boundary markers: The skill does not explicitly instruct the agent to use delimiters or ignore instructions found within fetched data, though it emphasizes "direct quotation only" and grounding claims in sources.
- Capability inventory: The skill is entirely instruction-based (Markdown) and does not include any executable scripts, file-writing tools, or its own network implementations.
- Sanitization: There is no evidence of specific sanitization or filtering logic applied to the external data fetched during the research phases.
Audit Metadata