wield
Pass
Audited by Gen Agent Trust Hub on Jul 27, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a robust safety framework that requires every automation to undergo a dry-run and hardening phase, specifically gating destructive operations like file deletion or email sending behind explicit confirmation or safe-mode logs.
- [SAFE]: Technical instructions in the integration guide emphasize security by design, warning against the interpolation of untrusted input into shell commands and recommending the use of discrete arguments (argv) to prevent injection attacks.
- [SAFE]: The documentation provides clear strategies for adhering to the principle of least privilege, advising the agent to minimize the TCC (Transparency, Consent, and Control) permission surface by limiting the number of target apps controlled.
- [SAFE]: Credential security is addressed through a 'Never' rule that forbids embedding secrets in plaintext, instead directing the agent to use the system Keychain via the 'security' utility.
- [SAFE]: The skill identifies and provides mitigation for indirect prompt injection risks by establishing clear operational boundaries and structured input/output protocols for multi-skill workflows.
Audit Metadata