learn-by-doing
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill design relies on processing user-provided project definitions and code modifications to generate learning modules and project files, which serves as an ingestion surface for external data.
- Ingestion points: Project goals provided by the user in Step 1 and code modifications submitted during the Stage 2 Modification Challenge.
- Boundary markers: The instructions do not define specific delimiters or guidelines to distinguish user-provided content from agent-generated instructions.
- Capability inventory: The skill involves writing code and documentation to the local file system within a project-specific directory structure.
- Sanitization: There are no explicit instructions for sanitizing, validating, or escaping user-provided project requirements or code changes before they are integrated into the project context.
Audit Metadata