research-lookup
Pass
Audited by Gen Agent Trust Hub on Jun 26, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8).
- Ingestion points: User-provided research queries enter the system via command-line arguments in
lookup.pyandresearch_lookup.pyand are directly interpolated into system prompts. - Boundary markers: Absent. The skill does not wrap user input in delimiters (e.g., XML tags or triple backticks) nor does it provide explicit instructions to the backend models to ignore potential instructions embedded within the user query.
- Capability inventory: The skill has the capability to perform network requests to external APIs (
api.parallel.aiandopenrouter.ai) and write research results to the local filesystem (sources/directory). - Sanitization: There is no evidence of input validation, escaping, or sanitization on the research queries before they are transmitted to external services.
Audit Metadata