draft-email-responses
Pass
Audited by Gen Agent Trust Hub on Jul 30, 2026
Risk Level: SAFEDATA_EXFILTRATIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill transmits sensitive email data including recipient addresses, thread IDs, subjects, and message bodies to an external n8n webhook endpoint via curl. This constitutes a potential data exposure path to an external service.
- [COMMAND_EXECUTION]: The skill utilizes shell commands to perform its tasks. It uses
curlto communicate with the n8n API andlsto inspect local file directories such asMarketing/brand-deals/andAgency/clients/for business context. - [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by ingesting untrusted email bodies via Gmail MCP tools. 1. Ingestion points: Email content retrieved through
mcp__claude_ai_Gmail__get_thread. 2. Boundary markers: Explicit instructions provided to treat scraped content as material for analysis only and to ignore any instructions contained within. 3. Capability inventory: Access to Gmail read tools and shell execution (curl). 4. Sanitization: Relies on natural language instructions for the agent rather than technical content sanitization.
Audit Metadata