stitch-vue-bootstrap-components

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). The skill explicitly instructs the agent to call Stitch MCP get_screen and download untrusted, user-generated Stitch HTML from htmlCode.downloadUrl (using scripts/fetch-stitch.sh) and then parse/interpret that HTML as part of its conversion workflow, so arbitrary third-party page content can influence tool use and decisions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly downloads Stitch screen HTML at runtime via the htmlCode.downloadUrl (fetched by scripts/fetch-stitch.sh using curl, e.g. a Stitch MCP / Google Cloud Storage URL such as https://storage.googleapis.com/...) and injects that fetched HTML into the conversion workflow so the remote content directly controls the agent's generated prompts/output.