brainstorm
Pass
Audited by Gen Agent Trust Hub on Oct 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill establishes an attack surface for indirect prompt injection by ingesting and processing untrusted external data.
- Ingestion points: External data enters the agent context via the
WebSearchtool during Phase 1 (Research First) inSKILL.md. - Boundary markers: The instructions in
SKILL.mdlack explicit boundary markers or delimiters when synthesizing research findings to protect the agent from potential instructions embedded in web content. - Capability inventory: The skill possesses the
Writetool capability, which is used to persist synthesized findings and analysis to the local file system in Phase 5 ofSKILL.md. - Sanitization: There is no mention of sanitization, filtering, or validation of the content retrieved from web searches before it is interpolated into the final brainstorming artifact and written to disk.
Audit Metadata