Skills
Claude Agent-style skill guides
AI Development
A working reference for embeddings, Python AI libraries, and core ML concepts — for RAG pipelines see RAG, for local model serving see Ollama, for AI-to-tool integration see MCP.
AI in Drupal
How an AI assistant is wired into Drupal: the direct-answer vs. agent-delegation flow, tool integration, and a working system prompt example.
AWS
A hands-on AWS CLI reference covering IAM, EC2, ECS, S3, RDS, Lambda, CodeBuild, monitoring, and cost optimization.
Claude Code Subagents
A practical guide to building a Claude Code subagent — choosing tools, model, scope, and the trigger description that makes the main session actually delegate to it.
Databases
A practical reference for connecting to, querying, indexing, and securing SQL and NoSQL databases.
Design Patterns
A one-page reference for applying SOLID principles and classic design patterns in Drupal and PHP projects.
DevOps Delivery Workflows
A trio of workflow skills — deploy, feature delivery, and tiny fixes — that scale the process to the size of the change.
Drupal Development & Operations
A set of skills covering Drupal 11 module development, security review, and DDEV operations.
Git
A working reference for everyday git — staging, committing, branching, merging, remotes, history, stashing, and how to undo changes without losing work.
Git Flow
A working reference for the Gitflow branching model — feature, release, and hotfix branches, the commands for each, and an honest note on when a simpler model fits better.
Git Worktrees
A skill for running parallel work on a git repository without conflicts, using isolated worktrees instead of stashing or switching branches.
Hugo
A practical reference for installing, configuring, and deploying Hugo static sites.
Kubernetes (K8S)
A practical skill with kubectl and Helm commands for managing pods, deployments, services, and troubleshooting a Kubernetes cluster.
LLM Attacks
Six classes of attacks on LLM-backed apps and concrete defensive principles: why both the model’s input and output are untrusted, how to handle output safely, how to limit tool agency, and how to isolate untrusted data in the prompt.
Model Context Protocol
A method reference for implementing or troubleshooting Model Context Protocol servers and clients — sessions, tools, resources, prompts, logging, and server info.
Node.js
A reference for managing Node.js versions, npm, Express.js, package.json, debugging, and the Vite/Next.js/Nodemon toolchain.
Ollama
A working reference for running Ollama locally: install commands, embedding model choices, client APIs, and a Docker Compose setup with optional GPU support.
OWASP ASVS Security Review
A skill for running OWASP ASVS v5.0 security verification, mapped to concrete Drupal 11 APIs and configuration.
PHP
A practical PHP developer reference covering coding standards, testing, frameworks, security hardening, performance tuning, and deployment.
Python
A practical Python reference covering core syntax, data science libraries, and workflow orchestration with runnable examples.
RAG
A working reference for building a RAG pipeline: chunking, dense/sparse/hybrid retrieval, reranking, an end-to-end example, and how to evaluate output quality.
System Prompts & AI Personas
A practical guide to authoring a reusable system prompt — the dimensions to define, a ‘You are…’ template, and a checklist for whether it will actually behave.
Temporal (Self-Hosted)
A practical ops reference for self-hosting Temporal on a VPS with Docker Compose — running the stack, the CLI, namespaces, workers, and single-host gotchas.
TypeScript
A practical TypeScript reference covering project setup, core and advanced types, the tsc CLI, module systems, decorators, error handling, and testing with Jest.