better-interface
Pass
Audited by Gen Agent Trust Hub on Aug 25, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill allows the agent to execute project-specific preview or test commands to verify interface behavior and visual consistency (Principle 12).
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection by ingesting and processing untrusted source code and project metadata from the local workspace environment.
- Ingestion points: The agent reads source files, user-facing strings, and internal signals such as analytics route lists (Principle 1, Principle 4).
- Boundary markers: Absent; the instructions do not implement specific delimiters to isolate external workspace content when dispatching data to sub-agent lanes.
- Capability inventory: The skill possesses capabilities to execute shell commands for verification (Principle 12) and modify workspace files in build mode (Principle 14).
- Sanitization: Absent; the skill relies on the core model's instruction-following capabilities rather than programmatic filtering of workspace content.
Audit Metadata