google-flights

Fail

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: CRITICALREMOTE_CODE_EXECUTIONCREDENTIALS_UNSAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [REMOTE_CODE_EXECUTION] (CRITICAL): The skill promotes the use of 'curl | bash' for installation from an untrusted domain (canifi.com).
  • Evidence: curl -sSL https://canifi.com/skills/google-flights/install.sh | bash
  • Evidence: curl -sSL https://canifi.com/install.sh | bash
  • This pattern bypasses all security checks and allows the remote server to execute arbitrary commands on the user's machine.
  • [CREDENTIALS_UNSAFE] (HIGH): The skill explicitly asks for and handles sensitive credentials, including SERVICE_PASSWORD and GOOGLE_EMAIL.
  • While it claims local storage via canifi-env, the presence of the critical RCE vulnerability mentioned above means these credentials can be easily exfiltrated by the installation scripts.
  • Evidence: The 'Authentication Flow' section hardcodes a specific email ember@emberbru.com, suggesting non-standard session handling.
  • [EXTERNAL_DOWNLOADS] (MEDIUM): The skill relies on resources and scripts hosted on canifi.com, which is not a trusted source according to security standards (e.g., it is not a verified organization like Google or Anthropic).
  • [COMMAND_EXECUTION] (HIGH): The skill setup requires the execution of shell commands that modify the local environment and potentially the filesystem (cp -r skills/google-flights ~/.canifi/skills/).
Recommendations
  • HIGH: Downloads and executes remote code from: https://canifi.com/skills/google-flights/install.sh, https://canifi.com/install.sh - DO NOT USE without thorough review
  • AI detected serious security threats
Audit Metadata
Risk Level
CRITICAL
Analyzed
Feb 17, 2026, 05:09 PM
Security Audit — agent-trust-hub — google-flights