stitch-vue-vant-components

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

• Third-party content exposure detected (high risk: 0.90). The skill's required workflow (SKILL.md Retrieval and Networking and Integration sections) explicitly instructs the agent to call stitch-mcp-get-screen and download user-provided design HTML and screenshots from htmlCode.downloadUrl (with scripts/fetch-stitch.sh) — i.e., ingesting external Stitch/URL-hosted, user-generated content that the agent must read and convert into code, so untrusted third-party content can influence subsequent actions.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill's runtime fetch (scripts/fetch-stitch.sh) uses curl to download the runtime-provided htmlCode.downloadUrl (e.g., the Stitch/GCS HTML URL passed into bash as "<htmlCode.downloadUrl>"), and that fetched HTML is parsed and injected into the agent's conversion process so it directly controls the generated prompts/instructions.