case-study-builder
Pass
Audited by Gen Agent Trust Hub on Nov 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input from the user (client details, project results, and quotes) which creates an indirect prompt injection surface. This is a common pattern for text-processing skills and is considered safe given the lack of dangerous capabilities.
- Ingestion points: Data is gathered through a structured 8-field questionnaire defined in
SKILL.md(Fields: Client, Before, Actions, After, Timeline, Quote, Naming, Use case). - Boundary markers: The skill uses a markdown-structured format and internal reasoning phases (Phases 1-4) to process data, though it does not explicitly use "ignore embedded instructions" delimiters.
- Capability inventory: The skill does not perform any subprocess calls, file system writes, or network operations. It is limited to text generation and internal self-critique logic.
- Sanitization: No explicit sanitization or escaping of user-provided content is performed before interpolation into the generation phases.
Audit Metadata