skills/wkentaro/agent-skills/ui-diff/Gen Agent Trust Hub

ui-diff

Pass

Audited by Gen Agent Trust Hub on Sep 24, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructions specify running python3 -m http.server to serve the generated UI diff from a temporary directory. This is used for local previewing of the HTML output when assets or iframes require an HTTP context.
  • [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect injection because it ingests and renders external HTML and CSS source code.
  • Ingestion points: The agent reads HTML/CSS sources provided by the user or from local servers to populate iframes or srcdoc attributes in SKILL.md.
  • Boundary markers: None specified in the instructions.
  • Capability inventory: The agent can write files to $TMPDIR and execute the Python built-in web server.
  • Sanitization: There are no explicit instructions to sanitize or escape the UI source code before rendering it in the comparison page.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 24, 2026, 06:40 AM