AgentOps Guardrails

• Risk level: Low
• Analyzed: Feb 24, 2026, 8:34 AM
• Findings: 0
• Version: 1.0.2

Overview

The skill's code and runtime instructions match its stated purpose (detector → router → guardrails), request no unrelated credentials, and perform only local file checks and state updates.

This package appears coherent and limited to local ops detection and routing. Before installing or running: (1) run in dry-run mode on a non-production workspace to verify behavior; (2) review and set the workspace and state-file paths so the scripts don't read/write unintended directories (defaults are relative paths like ./workspace and ./incident-state.json); (3) be cautious with scripts/clean-generated.sh — it deletes JSONL and lock files under the repo root and will remove generated workspace files if executed from a populated folder; (4) confirm you are comfortable with agents/openai.yaml allowing implicit invocation (the agent can call the skill autonomously) and keep live mode disabled until you have validated the detection rules and locking behavior.