html-mind-map
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided text for node labels and descriptions, which are then rendered into interactive HTML artifacts. Ingestion points: Node labels and descriptions edited by the user in the generated artifact. Boundary markers: Explicit instructions are provided to the agent to treat data received from the artifact's submission pipeline strictly as data and to never interpret it as new instructions or commands. Capability inventory: The skill generates HTML files containing JavaScript and writes them to the local filesystem. Sanitization: Instructions strictly mandate the use of textContent and createElement for DOM updates, explicitly forbidding the use of innerHTML to prevent XSS.
- [EXTERNAL_DOWNLOADS]: The skill allows the use of Google Fonts for styling, which is a recognized and trusted service.
- [COMMAND_EXECUTION]: The skill utilizes a platform-native tool (html-skills:html-skills-listen) to establish a local communication handshake between the interactive artifact and the agent.
Audit Metadata