visual-overlay
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
- [PROMPT_INJECTION] (LOW): Indirect Prompt Injection Surface. The skill grants the agent permissions to read and edit application source code using standard tools. While necessary for its functionality, this creates a vulnerability surface where the agent could be influenced by malicious content embedded in the files it modifies.\n- Ingestion points: Application source code and configuration files accessed via Read and Glob tools.\n- Boundary markers: Absent; no specific delimiters or instructions are provided to help the agent isolate user content from instructions.\n- Capability inventory: Filesystem write and modification access provided via Edit and Write tools.\n- Sanitization: Absent; the agent is expected to directly modify files without a sanitization layer.
Audit Metadata