chatwork-automation
Warn
Audited by Snyk on Feb 16, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).
- Potentially malicious external URL detected (high risk: 1.00). The skill requires adding the MCP server https://rube.app/mcp as a runtime endpoint and uses RUBE_SEARCH_TOOLS / RUBE_MULTI_EXECUTE_TOOL against it to fetch tool schemas and execute remote tools, meaning fetched content can directly control agent prompts/instructions and trigger remote code execution.
Audit Metadata