email-invoice-processor

Warn

Audited by Gen Agent Trust Hub on Sep 21, 2026

Risk Level: MEDIUMDATA_EXFILTRATIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONCREDENTIALS_UNSAFEDYNAMIC_EXECUTION
Full Analysis
  • [DATA_EXPOSURE]: The skill connects to the user's email via IMAP to read and filter messages, exposing sensitive communications to the agent's context.
  • [EXTERNAL_DOWNLOADS]: The skill extracts URLs from untrusted email bodies and automatically downloads files using requests or playwright.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted email content without sanitization, creating an attack surface for malicious instructions. (Ingestion: email subject/body/attachments; Capabilities: network/file-write/browser; Sanitization: None for content).
  • [COMMAND_EXECUTION]: The documentation provides shell commands for environment setup and script execution.
  • [CREDENTIALS_UNSAFE]: Email authorization codes are stored in plain text within a config.json file.
  • [DYNAMIC_EXECUTION]: The skill uses playwright to navigate and interact with untrusted websites to download invoices, which involves executing remote JavaScript code.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Sep 21, 2026, 03:33 AM