onepassword-environment-secrets

Warn

Audited by Gen Agent Trust Hub on Jul 31, 2026

Risk Level: MEDIUMCREDENTIALS_UNSAFECOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [CREDENTIALS_UNSAFE]: The skill instructions specify access to sensitive credential files located at ~/.config/op/service-account-token and ~/.config/op/environment-id to authenticate the agent with 1Password.\n- [COMMAND_EXECUTION]: The skill performs shell execution via the op command-line tool, including commands to read environment configurations and execute arbitrary logic within a subshell using op run --environment ... -- sh -c ....\n- [DATA_EXFILTRATION]: The skill processes highly sensitive environment variables (e.g., OPENAI_API_KEY) and 1Password Environments. The ingestion of these secrets into the agent context creates a potential exfiltration vector if instructions to suppress output are bypassed. \n
  • Ingestion points: 1Password Environment data read via op CLI (SKILL.md).\n
  • Boundary markers: The skill contains explicit rules to never print, reveal, or summarize secret values and uses redirection to /dev/null for verification commands (SKILL.md).\n
  • Capability inventory: Shell command execution (op, sh) and sensitive file system access (SKILL.md).\n
  • Sanitization: Suppression of command output using >/dev/null is suggested for verification steps.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jul 31, 2026, 01:34 PM
Security Audit — agent-trust-hub — onepassword-environment-secrets