ordercli

Warn

Audited by Gen Agent Trust Hub on Aug 1, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill automates the installation of an external binary ordercli from a third-party GitHub repository (github.com/steipete/ordercli) using Homebrew or Go.\n- [DATA_EXFILTRATION]: The skill provides instructions for the agent to access highly sensitive user data, including browser cookies and session profiles, to authenticate with food delivery services.\n
  • Evidence: Use of ordercli foodora cookies chrome --profile "Default" and similar commands to import browser state.\n- [CREDENTIALS_UNSAFE]: The skill handles user authentication secrets including passwords via standard input and bearer tokens stored in environment variables.\n
  • Evidence: Instructions for ordercli foodora login --password-stdin and references to DELIVEROO_BEARER_TOKEN.\n- [COMMAND_EXECUTION]: The skill relies on executing the ordercli utility to perform network-active actions such as placing orders and modifying account configurations.\n- [PROMPT_INJECTION]: The skill processes external data from food delivery APIs without defined boundary markers or sanitization, creating a potential surface for indirect prompt injection.\n
  • Ingestion points: Order history and active order status details.\n
  • Boundary markers: Absent.\n
  • Capability inventory: Shell command execution via ordercli (including reordering and account login).\n
  • Sanitization: Absent.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Aug 1, 2026, 08:32 AM
Security Audit — agent-trust-hub — ordercli