Claude Skills by mgajewskik

Senior-level Ansible guidance for production automation. Use when designing, creating, running, reviewing, or debugging Ansible projects: ansible-core execution, playbooks, roles, collections, custom modules/plugins, inventories, variable flow, EEs, AWX/AAP, PAH, EDA, air-gap workflows, testing, upgrades, performance, secrets, and large-scale rollouts.

Optional idea-shaping skill for when the user explicitly wants to brainstorm, compare approaches, pressure-test a concept, or turn a rough idea into a lightweight plan. Do not trigger for routine implementation, small edits, or cases where the user clearly wants direct execution.

Apply the Diátaxis documentation framework to learn the model, classify documentation, reformat existing docs, and draft new tutorials, how-to guides, reference, and explanation. Use when the user mentions Diátaxis, tutorial vs how-to, reference vs explanation, documentation quadrants, restructuring docs by user need, or writing docs in Diátaxis style.

Create senior-level deep research dossiers and roadmap companions. Use when the user asks for a dossier, senior research, deep research, in-depth research, mental models for a topic, senior perspective on a topic, how something actually works, ramp up on a topic, architectural deep dive, tradeoffs, failure modes, or what a senior would notice. Produces current-directory research-* and roadmap-* markdown artifacts, not a tutorial or short summary.

Create proper git commits from repository changes using Conventional Commits headers and a high-signal commit body that explains why the change was needed and what was done. Use when the user says commit this, make a commit, create a commit, draft a commit message from current changes, or asks for a proper git commit.

Unified GitLab skill for issues, merge requests, reviews, CI/CD, releases, and repository operations. Use whenever the user mentions GitLab, GitLab issues, merge requests/MRs, pipelines, releases, repo administration, or glab. Prefer the bundled Python CLI first for issue and MR workflows; if that path is unavailable, unsupported, or fails, fall back to glab CLI and load only the matching reference module.

Jira and optional Confluence integration for issue tracking and documentation. Use when the user mentions Jira tickets or issue keys such as PROJ-123, asks about sprints, boards, or backlogs, or wants to read, search, create, update, transition, comment on, or organize Jira issues and Confluence pages.

Senior-level Knative and OpenShift Serverless guidance for Serving, Eventing, Functions, autoscaling, scale-to-zero, CloudEvents, RabbitMQ/Kafka sources, Lambda migration, Harbor/OCI images, debugging, operations, and production rollout. Use when working with Knative Service, Revision, Route, KPA, activator, queue-proxy, Broker, Trigger, Source, Sink, kn func, OpenShift Serverless, Kourier, eventing-rabbitmq, Knative Kafka, or serverless workloads on Kubernetes/OpenShift.

Create and render Mermaid diagrams with correct syntax and the mmdc CLI. Use when the user asks to "create a diagram", "draw a flowchart", "make a sequence diagram", "render mermaid", "generate a chart", "visualize architecture", "create an ER diagram", mentions "mermaid", asks for any visual diagram (flowchart, sequence, class, state, ER, gantt, git graph, pie, mind map, timeline, quadrant, sankey, XY chart, block, packet, kanban, architecture, radar, treemap, C4, user journey, requirement d...

Consult AI models (GPT, Gemini, Grok, Opus) for perspectives on questions. Use when user says "ask gpt", "ask gemini", "ask grok", "ask opus", "ask both", "ask all three", "ask all four", "ask gpt and gemini", "consult opus", or wants model opinions on decisions. Spawns ONLY the models explicitly requested.

Create and configure OpenCode customizations - commands, agents, prompts, modes, and AGENTS.md. Use when user says "create command", "add agent", "opencode prompt", "create mode", "edit AGENTS.md", "customize opencode", "opencode config", or needs guidance on OpenCode-specific configuration patterns.

Senior-level OpenShift and OKD guidance for installing, operating, debugging, upgrading, and securing clusters. Use when working with OpenShift Container Platform, OKD, MicroShift, oc, openshift-install, CVO, MCO, OLM, SCCs, Routes, OVN-Kubernetes, RHCOS/SCOS, disconnected or air-gapped installs, OperatorHub, ODF, OpenShift Virtualization, monitoring, logging, GitOps, Pipelines, Service Mesh, etcd, certificates, or day-2 cluster incidents.

Create and render PlantUML diagrams with correct syntax and the plantuml CLI. Use when the user asks to "create a puml diagram", "draw a plantuml diagram", "generate puml", "visualize with plantuml", mentions "puml" or "plantuml", or asks for PlantUML-specific diagrams (sequence, class, component, deployment, activity, ER, state, mindmap, gantt, C4, AWS architecture, K8s, use case, timing, nwdiag, network diagram, wireframe, salt mockup, archimate, EBNF grammar, regex diagram, ditaa). Do NOT ...

Senior-level Proxmox VE guidance for VM creation, templates, storage, ZFS, Ceph, networking, clusters, HA, PBS backups, debugging, upgrades, security, and production/homelab operations. Use when working with Proxmox, PVE, Proxmox VE, qm, pct, pvesm, pvecm, pmxcfs, HA manager, Proxmox Backup Server, VM migration, Proxmox incidents, or Ceph/ZFS/Corosync/VLAN bridges in a Proxmox VE context.

Create effective PRDs (Product Requirements Documents) for Ralph loops / Ralph Wiggum autonomous AI coding agents. Use whenever the user wants to create a PRD, task list, or specification for use with Ralph loops, Claude Code loops, Codex loops, or any autonomous agent loop. Also use when the user says 'create a PRD', 'write a task list for ralph', 'break this down into stories', 'plan this feature for autonomous coding', 'help me write acceptance criteria', 'prepare this for a ralph loop', o...

Senior-level RHEL-family Linux operations. Use when running, debugging, hardening, patching, installing, upgrading, or operating Red Hat Enterprise Linux, Rocky Linux, AlmaLinux, CentOS Stream, Fedora-as-upstream, or related enterprise Linux hosts: systemd, RPM/DNF, SELinux, NetworkManager, firewalld, storage, kernel/kdump, FIPS/STIG, Satellite, IdM, Podman, bootc, air-gapped fleets.

Create and manage bash task runner scripts (Makefile alternative). Use when user says "create run script", "add to run script", "add task to run", "create helper in run script", "run script task", or asks for a bash task runner.

Design, create, evaluate, and improve Claude Skills with structure matched to the task. Use when creating a new skill, auditing or refactoring an existing skill, deciding how a skill should be structured, choosing between Simple/Complex/Lightweight archetypes, adding evals or validation patterns, optimizing a skill description, or packaging a skill for reuse. Covers discovery interviews, archetype selection, structural patterns, evaluation loops, benchmarking, and trigger-tuning.

Use when writing, reviewing, or debugging Terraform/OpenTofu modules, tests, CI, scans, or state ops — diagnoses failure mode (identity churn, secrets, blast radius, CI drift, state corruption) with version-aware guards.

Draft, rewrite, critique, copyedit, or coach prose for humans to read: email, chat, LinkedIn, blogs, docs, READMEs, PRs, commits, issues, ADRs, design docs, incident updates, postmortems, status updates, executive summaries, instructions, and general writing. Use when the user asks for writing, editing, feedback, clearer/concise prose, Strunk-style polish, anti-AI-slop cleanup, or reusable writing coaching. Preserve meaning, evidence, constraints, and voice while improving reader-task fit.