analyse-design
Pass
Audited by Gen Agent Trust Hub on Mar 2, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: Indirect prompt injection surface detected through ingestion of untrusted codebase data.
- Ingestion points: Reads local codebase files (CSS, JS, JSON) and screenshots using Read, Grep, and Glob tools.
- Boundary markers: Not present; the instructions do not include delimiters or warnings to disregard instructions within the analyzed data.
- Capability inventory: Read, Write, Edit, Grep, and Glob.
- Sanitization: Not present; files are analyzed in their raw form.
Audit Metadata