research
Pass
Audited by Gen Agent Trust Hub on Mar 9, 2026
Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [DATA_EXFILTRATION]: Sends user-provided research topics to an external API endpoint (https://api.inspiro.top/research). This transmits potentially sensitive query data to a third-party domain not associated with the vendor or trusted services.
- [PROMPT_INJECTION]: Vulnerable to indirect prompt injection via the research data ingestion surface. 1. Ingestion points: Results retrieved from the Inspiro Research API in scripts/research.sh. 2. Boundary markers: None; the skill does not use delimiters or warnings to prevent the agent from following instructions embedded in research results. 3. Capability inventory: Execution of shell commands and writing to local files. 4. Sanitization: No sanitization is performed on the data returned from the API.
- [COMMAND_EXECUTION]: Utilizes system commands curl and jq within scripts/research.sh to facilitate API communication and result parsing. While common for such tasks, these operations handle external data without additional security layers.
Audit Metadata