youtube-automation

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

• Third-party content exposure detected (high risk: 0.75). The required workflow is driven by Composio YouTube tool calls (e.g., search/list/comment threads), which at runtime can return outsider-authored free text such as video titles/descriptions and especially comment bodies from other users, and that returned text is then ingested into the agent’s LLM context via the tool results.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill instructs the agent to add and use the MCP endpoint https://rube.app/mcp at runtime (via RUBE_SEARCH_TOOLS) to fetch tool schemas that directly control available tool calls and agent behavior, so this external URL is a required runtime dependency that can influence prompts/instructions.