creativeclaw

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

• Third-party content exposure detected (high risk: 0.80). The skill's workflow docs explicitly require fetching and interpreting external web content (e.g., "website extraction" in references/workflows/brand-theme.md and "reference-image collection (existing files, URLs, or import_media drop-in)" in references/workflows/template-creation.md), and those imported public URLs and site contents are used to build themes/prompts that materially drive generation decisions, so it clearly ingests untrusted third-party content that could carry indirect prompt injection.

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

• Potentially malicious external URL detected (high risk: 1.00). The edit-video install steps include a runtime command that downloads-and-executes remote code (curl -LsSf https://astral.sh/uv/install.sh | sh) and remote binary downloads (e.g. https://evermeet.cx/ffmpeg/...), which are fetched during setup and execute remote code as required dependencies for the edit-video workflow.