cold-email-verifier
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes lead data from a user-provided CSV file, creating a surface for indirect prompt injection if the agent reads and follows instructions embedded within the file content. 1. Ingestion points: The scripts/email_verifier.py script reads the input file using pandas.read_csv(). 2. Boundary markers: No explicit delimiters or instructions are used to prevent the agent from following commands found within the CSV data. 3. Capability inventory: The skill utilizes subprocess.run for CLI execution, requests for network access to external APIs, and pandas for file system write operations. 4. Sanitization: While the script performs regex-based cleaning on names to generate email permutations, it does not sanitize the raw input data before it is processed by the agent.
- [COMMAND_EXECUTION]: The skill executes external CLI commands to verify emails using the Reacher engine. 1. Evidence: In scripts/email_verifier.py, the verify_reacher_cli function calls subprocess.run using a list of arguments (['check_if_email_exists', email]). This is a secure implementation that avoids shell injection vulnerabilities.
- [DATA_EXFILTRATION]: The skill transmits data to external services as part of its core functionality. 1. Evidence: It sends company names to Clearbit's Autocomplete API and email addresses to ValidEmail.co for deliverability verification. These are well-known or primary service providers for the skill's stated purpose.
Audit Metadata