preflight-checks

Pass

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill's operational contract is strictly limited to read-only inspection (R0-R1 risk scope). It is explicitly forbidden from performing destructive, remote, or history-rewriting actions, ensuring no unauthorized changes are made to the user's repository.
  • [CREDENTIALS_UNSAFE]: The skill incorporates defensive logic to identify secret-like values in staged files or environment files (e.g., .env). It mandates that any detected secrets must be redacted in the output and the task must be routed to a dedicated security scanning skill, preventing accidental credential exposure.
  • [COMMAND_EXECUTION]: Tool usage is restricted to standard Git diagnostic commands such as git status and git diff. The skill configuration prevents the execution of long-running or destructive shell commands and enforces user confirmation for any operation involving index changes.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 22, 2026, 04:19 AM
Security Audit — agent-trust-hub — preflight-checks