research-lookup

Pass

Audited by Gen Agent Trust Hub on Feb 21, 2026

Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [DATA_EXFILTRATION] (LOW): The skill is designed to transmit user-provided research queries to the OpenRouter API (openrouter.ai). Although this is the primary functionality of the tool, the destination domain is not included in the trusted whitelist, representing a standard network exposure for external API interactions.
  • [PROMPT_INJECTION] (LOW): The skill presents an indirect prompt injection vulnerability surface. 1. Ingestion points: Untrusted user input is ingested via command-line arguments (sys.argv) in lookup.py. 2. Boundary markers: The format_response function in lookup.py does not implement delimiters or instructions for the agent to ignore potentially malicious content embedded in the research results. 3. Capability inventory: The skill possesses the capability to perform external network operations and return data to the agent context. 4. Sanitization: There is no sanitization or validation of the content returned by the OpenRouter API before it is formatted and displayed to the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 21, 2026, 04:21 AM