rsturla-skills

Diagnose and troubleshoot OpenShift cluster and workload issues. Pod failures, networking, storage, operators, SCCs, and node problems. Use when the user has an OpenShift issue, mentions oc commands, pod crashes, CrashLoopBackOff, ImagePullBackOff, routes not working, or cluster problems — even if they just say "my pod won't start" or "the app is broken."

Build TypeScript AI agents with eval framework, provider abstraction, and production guardrails. Uses Vercel AI SDK for model-agnostic tool-use loops, Zod for structured outputs, Vitest for evals, and OpenTelemetry for observability. Covers prompt caching, extended thinking, RAG pipelines, structured output enforcement, tool annotations, seed tool calls, toolcall regression evals, and duplicate call detection. Use when the user wants to build an agent, create an AI assistant, implement tool-use, or scaffold an agentic system — even if they just say "build a bot", "add AI to this", or "make it agentic."

Comprehensive code review using multiple specialized agents. Checks architecture, error handling, naming, testing, and style against project conventions. Use when the user asks for a code review, says "review this", "check my code", or wants feedback on changes — even if they just say "what do you think?" about a diff.

Generate or improve git commit messages from staged changes using Claude on Vertex AI. Conventional Commits format, validates line lengths, retries on failure. Use when the user wants to write, generate, or improve a commit message, asks for help with git commits, or invokes /commit-message — even if they just say "commit this" or "what should the commit say."

Audit project dependencies for known vulnerabilities, license issues, and supply chain risks. Supports Go, Rust, and Python. Use when the user changes dependencies, asks about vulnerability scanning, mentions CVEs in deps, or wants a security check — even if they just say "are my deps safe?" or "check for vulnerabilities."