skills/malue-ai/dazee-small/ordercli/Gen Agent Trust Hub

ordercli

Warn

Audited by Gen Agent Trust Hub on Mar 28, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSDATA_EXFILTRATIONCREDENTIALS_UNSAFECOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill facilitates the installation of the 'ordercli' binary from third-party repositories.
  • Source: GitHub (github.com/steipete/ordercli) and Homebrew tap (steipete/tap/ordercli).
  • [DATA_EXFILTRATION]: The skill accesses and processes sensitive local browser data to manage delivery service sessions.
  • Evidence: Commands such as 'ordercli foodora cookies chrome' and 'ordercli foodora session chrome' interact with local Chrome browser profiles and cookies.
  • Evidence: The tool manages persistent session data in '$HOME/Library/Application Support/ordercli/browser-profile'.
  • [CREDENTIALS_UNSAFE]: The documentation describes handling sensitive authentication material in plaintext or environment variables.
  • Evidence: Supports password entry via '--password-stdin'.
  • Evidence: References use of sensitive tokens such as 'DELIVEROO_BEARER_TOKEN' and 'DELIVEROO_COOKIE'.
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute the 'ordercli' binary with various parameters on the host system.
  • [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection by processing external data from food delivery APIs.
  • Ingestion points: Order lists and history details fetched from Foodora and Deliveroo APIs.
  • Boundary markers: Absent; the skill does not use delimiters to isolate API-provided data from agent instructions.
  • Capability inventory: Command execution (ordercli), network access, and sensitive file access (browser profiles).
  • Sanitization: No sanitization of API-provided data is documented.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 28, 2026, 02:07 AM