rafaelh

rafaelh/skills

Plugin Claude Code

Rafe Hart's Claude Code plugins.

rafaelh/skills

Skill Claude CodeCodex

Build, run, smoke-test, and drive the skills marketplace and its bundled agent tools. Use when asked to run this repo, start or exercise the skill-audit / agent-tool-builder scripts, verify a change to a bundled script or SKILL.md actually works, check Python/TypeScript parity, run the tests, lint, typecheck, or…

rafaelh/skills

Skill Claude CodeCodex

Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.

rafaelh/skills

Agent Claude Code

Analyze blind comparison results to understand WHY the winner won and generate improvement suggestions.

rafaelh/skills

Agent Claude Code

Compare two outputs WITHOUT knowing which skill produced them.

rafaelh/skills

Agent Claude Code

Evaluate expectations against an execution transcript and outputs.

rafaelh/skills

Instructions file

Claude Code instructions for rafaelh/skills, covering claude.md, what this repo is, commands, architecture and skill layout is a contract, not a convention.

rafaelh/skills

Agent Claude Code

Run the full build pipeline and report pass/fail.

rafaelh/skills

Agent Claude Code

Audit content for consistency and quality issues.

rafaelh/skills

Agent Claude Code

Check whether documentation matches the current code.

rafaelh/skills

Agent Claude Code

Generate test scaffolding for untested modules.

rafaelh/skills

Skill Claude CodeCodex

Add a new transform type to the pipeline. Use when the user wants to add a new data transformation (e.g. join, aggregate, deduplicate) to the formwork pipeline.

rafaelh/skills

Skill Claude CodeCodex

Refactor code for clarity, maintainability, and reduced complexity. Use when code works but is harder to read, maintain, or extend than it should be.

rafaelh/skills

Skill Claude CodeCodex

Use this skill when the user wants to build features or fix bugs using TDD (test-driven development), mentions red-green-refactor, or asks for test-first development.

rafaelh/skills

Plugin Claude Code

Build and audit Python scripts intended to be called by AI agents as tools.

rafaelh/skills

Skill Claude CodeCodex

Build or audit Python scripts called by AI agents as tools. BUILD mode: when the user asks you to write, scaffold, or improve a script an agent will invoke — CLI tools, automation helpers, data-fetching or query utilities, anything where machine-readable output, single-turn completeness, predictable exit codes, and…

rafaelh/skills

Plugin Claude Code

Audit code against ADRs and layer separation, find deepening opportunities, write ADRs and CONTEXT.md domain docs, and design module interfaces.

rafaelh/skills

Skill Claude CodeCodex

Architecture work across five modes — (1) audit code for ADR and separation-of-concerns violations, (2) find deepening / refactoring opportunities, (3) write or update ADRs, (4) write or update CONTEXT.md / CONTEXT-MAP.md domain docs, (5) design module interfaces. Informed by the project's domain glossary (CONTEXT.md…

rafaelh/skills

Plugin Claude Code

Use the Azure CLI to inspect, manage, deploy, or troubleshoot Azure resources and investigate Azure security questions.

rafaelh/skills

Skill Claude CodeCodex

Use the Azure CLI (az) to inspect, manage, deploy, or troubleshoot Azure resources, and to investigate Azure security questions. Trigger whenever the user asks to create, configure, list, audit, or interact with anything in Azure — App Service and other PaaS, storage, networking, identity, RBAC — even if they do not…

rafaelh/skills

Plugin Claude Code

Audit a repo for work that should run on a cheaper model or in an isolated context, write the subagent definitions, and wire them into CLAUDE.md and the repo's skills.

rafaelh/skills

Skill Claude CodeCodex

Audit a repository for work that should run on a cheaper model or in an isolated context, then write .claude/agents/.md subagent definitions fitted to that repo's toolchain and wire them into its CLAUDE.md and skill workflows so they actually get invoked instead of sitting unused. Use when the user wants to cut token…

rafaelh/skills

Plugin Claude Code

Read, author, audit, and refactor Meta Bind syntax in an Obsidian vault — input fields, view fields, buttons, embeds, and the Meta Bind / JS Engine APIs.

rafaelh/skills

Skill Claude CodeCodex

Use this skill when reading, authoring, auditing, or refactoring Meta Bind syntax in an Obsidian vault — input fields, view fields, buttons, embeds, and the Meta Bind / JS Engine APIs. Trigger on INPUT[, VIEW[, BUTTON[, meta-bind-button, js-engine, "Meta Bind", "metabind", or the obsidian-meta-bind-plugin folder. Also…