html-playground
Pass
Audited by Gen Agent Trust Hub on Aug 10, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill requires the agent to open the generated HTML file in a browser to interact with controls and verify functionality. This is a legitimate workflow step for tool building.
- [COMMAND_EXECUTION]: The skill generates interactive code (HTML/JS) from local templates. This generation process is contained and does not execute untrusted remote code.
- [PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it ingests user context to build tools that output prompts. This is a characteristic of the skill's purpose and lacks malicious intent.
- Ingestion points: User requirements and product context in SKILL.md.
- Boundary markers: None specified in the templates.
- Capability inventory: File writing and browser interaction.
- Sanitization: No explicit sanitization of user-provided input is performed.
Audit Metadata