starting-a-new-project
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from both the user's initial project request and external websites to generate local project documentation.
- Ingestion points: Data enters the agent's context through the user's initial project idea and external web content retrieved via the
web_searchandfetch_contenttools. - Boundary markers: The skill does not explicitly define technical sanitization or boundary markers (such as delimiters or 'ignore' instructions) for fetched content, although the flow includes a mandatory human review and confirmation step before finalization.
- Capability inventory: The agent has the capability to write to the local file system using the
spec_createandedittools. - Sanitization: No specific technical sanitization or filtering logic is described for the fetched content before it is proposed to the user or saved to the workspace.
Audit Metadata