break
Pass
Audited by Gen Agent Trust Hub on Aug 25, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill analyzes local component code (props, slots, states, and data) to infer test scenarios. Malicious instructions embedded within component source code could potentially influence the agent during the analysis or rendering phase.
- Ingestion points: Component source files (props, slots, states, data) as described in SKILL.md Step 2.
- Boundary markers: The instructions do not define delimiters or provide warnings to ignore embedded instructions within the source code being analyzed.
- Capability inventory: The skill is authorized to write files (harness page), interact with local dev servers, and perform browser automation/screenshots.
- Sanitization: No sanitization or validation of the ingested component code is performed before it is processed or rendered.
- [COMMAND_EXECUTION]: The skill requires the agent to interact with the project's development environment, including creating routes in dev servers or writing and executing standalone HTML/JS files to render components. This involves executing local commands and project-specific code.
- [DYNAMIC_EXECUTION]: The agent dynamically generates a "harness page" that imports and executes real components from the project. This runtime script generation and execution is a core part of the skill's functionality but presents an inherent security risk if the imported code is malicious.
Audit Metadata