brainstorming
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill instructions and configuration do not contain any malicious command execution, network requests, or credential handling. It follows security best practices by requiring explicit invocation.
- [INDIRECT_PROMPT_INJECTION]: The skill implements a workflow that takes user input to generate subagent prompts, which is a standard pattern for this type of orchestration but creates a surface for indirect prompt injection.
- Ingestion points: The
$brainstormingvariable inSKILL.mdcaptures user input which is then used to frame subagent tasks. - Boundary markers: The instructions do not specify explicit delimiters or "ignore instructions" markers when passing the topic to subagents.
- Capability inventory: The skill utilizes subagent spawning via
fork_turns: "none"but does not involve file system writes, network calls, or subprocess execution. - Sanitization: No explicit sanitization or filtering of the
$brainstorminginput is performed prior to prompt generation.
Audit Metadata