advise
Pass
Audited by Gen Agent Trust Hub on Jul 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes shell commands via the
uv runinterface to execute a local tool namedresearchkit. This pattern incorporates user-supplied strings as command-line arguments. - [PROMPT_INJECTION]: The skill serves as a multiplexer for user queries, sending them to various independent AI models.
- Ingestion points: User input provided via the
<question>argument in theSKILL.mdusage examples. - Boundary markers: No specific delimiters or "ignore instructions" wrappers are visible in the provided command templates.
- Capability inventory: The skill executes external CLI tools (Claude, Codex, etc.) which have broad code-generation and analysis capabilities.
- Sanitization: There is no documented validation or sanitization of the question before it is passed to the downstream agents.
Audit Metadata