stigg
Pass
Audited by Gen Agent Trust Hub on Aug 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill mandates a 'Search-First Rule', requiring the agent to fetch and process information from external sources prior to generating code. While this ensures accuracy and prevents the use of stale snippets, it creates a surface where external documentation could influence agent behavior.
- Ingestion points: The agent is instructed to fetch content from
docs.stigg.io(includingllms.txt),mcp.stigg.io, andwidgets.stigg.ioto gather integration context. - Boundary markers: The instructions do not specify explicit delimiters or 'ignore' instructions for the external content fetched during these operations.
- Capability inventory: The skill coordinates sub-skills (such as
stigg-apiandstigg-mcp) that perform API interactions, network requests, and code generation based on the retrieved documentation. - Sanitization: No explicit sanitization or validation of the remote documentation content is described, though the source is the official vendor domain.
Audit Metadata