html-mind-map

Pass

Audited by Gen Agent Trust Hub on May 18, 2026

Risk Level: SAFE
Full Analysis
  • [DATA_EXFILTRATION]: The skill implements a submission pipeline to return user-generated mind map data to the agent. This occurs via a local POST request to a loopback address (127.0.0.1) or through the clipboard. These mechanisms are documented, session-scoped, and intended for local development workflows, posing no exfiltration risk to remote actors.
  • [INDIRECT_PROMPT_INJECTION]: The skill facilitates an interactive loop where user input from the browser-based mind map is ingested by the agent.
  • Ingestion points: Data enters the agent's context through a POST request to the local listener or a clipboard paste containing a JSON payload.
  • Boundary markers: The skill requires a standardized JSON envelope with 'skill', 'kind', and 'data' fields to structure the feedback.
  • Capability inventory: The skill is capable of writing local HTML files and invoking the html-skills:html-skills-listen tool.
  • Sanitization: The instructions explicitly mandate the use of textContent and appendChild over innerHTML for all variable data to mitigate Cross-Site Scripting (XSS) risks within the generated artifact.
  • [COMMAND_EXECUTION]: The skill directs the agent to use the html-skills:html-skills-listen tool to facilitate its interactive features. This is a standard tool invocation within the platform's execution environment.
Audit Metadata
Risk Level
SAFE
Analyzed
May 18, 2026, 12:58 PM
Security Audit — agent-trust-hub — html-mind-map