level-up
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill establishes an attack surface by ingesting user answers and saving them verbatim into project files (
notes/learning/user-knowledge.mdandnotes/learning/LEARNING-PLAN.md) intended for consumption by future agents. - Ingestion points: User-provided answers during the 7-question assessment phase.
- Boundary markers: The instructions do not specify the use of delimiters or sanitization wrappers when writing user content to the persistent storage files.
- Capability inventory: The skill utilizes file system read/write capabilities for project documentation and source code reading for question context.
- Sanitization: No sanitization or escaping mechanisms are described; answers are stored as raw text.
Audit Metadata