skills/vdesjardins/nix-config/flux/Gen Agent Trust Hub

flux

Pass

Audited by Gen Agent Trust Hub on Feb 19, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • Indirect Prompt Injection (LOW): The skill interacts with Kubernetes cluster data and has administrative capabilities, creating a surface for indirect prompt injection. Ingestion points: kubectl describe, kubectl logs in TROUBLESHOOTING-WORKFLOWS.md. Capability inventory: kubectl apply, kubectl patch, flux reconcile in COMMON-COMMANDS.md. Boundary markers: None. Sanitization: None.
  • Suspect Image Reference (LOW): Use of alpine/fluxcd:latest in troubleshooting examples is non-standard, as the official image is fluxcd/flux-cli.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 19, 2026, 12:32 PM