stigg-governance
Pass
Audited by Gen Agent Trust Hub on Nov 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to elicit arbitrary, customer-defined model definitions (such as names for teams, projects, and cost centers) from users and utilize these inputs within SDK code executions via the
executetool. This creates a potential attack surface for indirect prompt injection if malicious strings are provided and processed without rigorous validation. - Ingestion points: User-provided inputs regarding unit names, nesting hierarchy, and attribution keys as described in the 'Elicit the Model' section of
SKILL.md. - Boundary markers: No explicit instructions or delimiters are provided to the agent for separating user-supplied strings from the SDK code templates during interpolation.
- Capability inventory: The skill enables a wide range of administrative operations (upserting, archiving, and querying) via the
@stigg/typescriptSDK using theexecutetool. - Sanitization: There are no specific instructions for sanitizing, escaping, or validating user-provided strings before they are used in the SDK code blocks.
Audit Metadata