multi-source-search
Pass
Audited by Gen Agent Trust Hub on Aug 26, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a robust evidence diversity model, explicitly instructing the agent to treat all retrieved content as untrusted input and to ignore any instructions embedded within it.- [SAFE]: The
validate_report.pyscript is a purely local validation utility that checks JSON structure and consistency without performing network operations or dynamic code execution.- [COMMAND_EXECUTION]: The skill uses a subprocess call to runpython3 scripts/validate_report.py research-report.json. This is a standard part of the workflow for validating locally generated research data and does not incorporate untrusted external input into the command line.- [EXTERNAL_DOWNLOADS]: The skill references SandBase search tools (Tavily, Exa, Scholar, Cloudsway) for multi-source coverage. These are recognized external search providers and are used in accordance with the skill's primary research purpose.
Audit Metadata