coda-automation

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly fetches and reads user-generated Coda content as part of its required workflows (e.g., CODA_RESOLVE_BROWSER_LINK, CODA_LIST_PAGES, CODA_GET_A_PAGE, CODA_LIST_TABLE_ROWS, and CODA_BEGIN_CONTENT_EXPORT) and then performs actions (search, upsert rows, push buttons, publish) based on that content, so untrusted/public Coda docs/pages could inject instructions that influence agent behavior.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly requires adding and using the Rube MCP endpoint https://rube.app/mcp and calls RUBE_SEARCH_TOOLS at runtime to fetch tool schemas, so that external URL is used during runtime and can directly influence the agent's prompts/instructions and behavior.