Quick Reference Card
| What I Want to Do | What to Type | Notes |
|---|---|---|
| Start my day | /morning | Pulls Todoist + Notion, suggests plan |
| Capture rambling thoughts | /brain-dump | Then paste or dictate |
| Create a LinkedIn post | /linkedin | Describe the topic or story |
| Check WP site before launch | /deploy-check | Run in a WordPress project directory |
| Check Next.js before deploying | /deploy-nextjs | Run in a Next.js project directory |
| Run accessibility audit | /a11y-audit | Then provide the website URL |
| End my session | /handoff | Generates summary for next session |
| Commit code changes | /commit | Creates a git commit with message |
| Commit + push + open PR | /commit-push-pr | Full git workflow in one command |
| Plan a complex task | /writing-plans | Before touching code on big features |
| Debug something broken | /systematic-debugging | Structured debugging, not guesswork |
| Review code quality | /requesting-code-review | After finishing a feature |
| Verify before claiming done | /verification-before-completion | Run this before saying "it's done" |
| Brainstorm before building | /brainstorming | Explore requirements first |
| Search past conversations | /search-conversations | Find what you discussed before |
| Clean up git branches | /clean_gone | Removes branches deleted on remote |
| Look up library docs | Ask Claude to use context7 | "Check the WordPress REST API docs using context7" |
| Scrape a webpage | /firecrawl-cli | Firecrawl handles all web operations |
| Generate an image | /generate | Describe what you want |
| Create interactive playground | /playground | Visual configurator for any concept |
| Build a polished UI | /frontend-design | Production-grade web interfaces |
| Get automation suggestions | /claude-automation-recommender | Recommends skills/agents for your task |
| Fix VS Code bash path error | /fix-vscode-bash-path | After VS Code extension updates |
| Browse the web directly | /browsing | Chrome DevTools Protocol control |
My Environment
Machine
| Component | Detail |
|---|---|
| OS | Windows 11 Pro (10.0.22621) |
| IDE | VS Code with Claude Code extension (Anthropic.claude-code) |
| Shell | Git Bash (C:\Program Files\Git\bin\bash.exe) |
| Node.js | v22.18.0, npm 11.6.4 |
| PHP | 8.2.12 (XAMPP) — upgrade to 8.4+ recommended for WP 6.9 |
| Python | 3.13 (C:\Python313\python.exe) |
| Formatter | Prettier (JS/TS), Intelephense (PHP) |
| Terminal | PowerShell (default), Git Bash available |
SSH Access
| Alias | What | Host | User | Port |
|---|---|---|---|---|
ssh hd | Hostinger web hosting (WordPress sites) | 141.136.39.73 | u945238940 | 65002 |
ssh vps | Hostinger KVM 2 VPS (n8n + OpenClaw Docker) | 72.62.212.169 | root | 22 |
SSH config: C:\Users\Bean\.ssh\config
Hosting
| Service | What It Hosts | Domain |
|---|---|---|
| Hostinger Cloud Startup | WordPress sites (dev + client) | palestine-lives.org (dev), various client domains |
| Hostinger KVM 2 VPS | n8n (Docker) + OpenClaw (Docker) | smallgiantsstudio.cloud, n8n.smallgiantsstudio.cloud |
| Vercel | Small Giants Studio website (Next.js) | smallgiantsstudio.co.uk |
| Cloudflare | DNS registrar + proxy | smallgiantsstudio.co.uk DNS |
| Supabase | Booking system database (eu-west-2, free tier) | Project ref: wimrjgrujprvwbsewqrq |
VS Code Claude Code Settings
| Setting | Value |
|---|---|
| Permission mode | bypassPermissions (dangerous mode enabled) |
| Preferred location | Panel |
| Model | Default (Opus 4.6) |
| Git Bash path | C:\Program Files\Git\bin\bash.exe |
| Agent teams | Enabled (experimental) |
My Projects
| Project | Stack | Status | Brief |
|---|---|---|---|
| HelpingDoctors EHR Pro | WordPress Multisite, PHP, ACF Pro, Ultimate Member | ~70% | Medical records system. HIPAA/GDPR compliant, ~70 DB tables, 27 medical roles, AES-256-GCM encryption, FHIR-compatible |
| Small Giants Studio | Next.js 16, React 19, Tailwind v4, TypeScript | LIVE | Business website at smallgiantsstudio.co.uk. Deployed on Vercel |
| SGS WordPress Framework | Custom block theme + plugins (PHP, JS, @wordpress/scripts) | In dev | sgs-theme, sgs-blocks (22+ blocks), sgs-booking, sgs-client-notes. Dev site: palestine-lives.org |
| Booking System | Next.js, Supabase, Drizzle, tRPC | Early | Standalone booking app with magic link auth |
| Mosque Timetable Plugin | WordPress, ACF, custom post types, shortcodes | Done | Prayer timetable display plugin |
| All Purpose Maintenance | n8n automation | Planned | Email intake, Asana, Xero, Tooltime integration |
My Daily Workflow
Morning Routine
Open VS Code terminal
Start Claude Code in whatever project you'll work on first.
Type: /morning
Claude pulls your Todoist tasks, checks Notion, and suggests a plan for the day.
Pick your first task and go
Start with the easiest quick win to build momentum (ADHD tip: small victories first).
During the Day
Switching Projects
Just tell Claude which project you're working on. It adjusts automatically.
Capturing Ideas
Type /brain-dump and talk or type freely. Claude structures it into actions, ideas, decisions, and worries.
If You're Stuck
Type /systematic-debugging for structured problem-solving. Don't guess — follow the process.
If You're Overthinking
Claude will call you out. It's in the rules. Listen to it.
End of Day
Type: /handoff
Claude generates a summary of everything done, what's pending, what's next.
Save the summary
Copy to Notion or leave in conversation history. Next session, Claude can pick up where you left off via episodic memory.
Weekly (Thursday/Friday)
Small Giants Studio Focus
Claude will remind you about business development. Review grant applications, client follow-ups, and networking. Check project status in Notion.
My Rules
What rules are
Markdown files in ~/.claude/rules/ that Claude loads automatically at the start of every conversation. They act as permanent instructions — you never need to repeat them. Some rules have paths: in their frontmatter, which means they only activate when working on matching file types.
All files uk-english.md
Enforces British English spelling everywhere: colour, behaviour, analyse, organisation, etc. Includes a full substitution table for common US/UK differences plus medical terms (paediatric, anaesthetic, haematology).
Exceptions: CSS properties, third-party library names, and existing database fields are kept as-is.
All code code-quality.md
Enforces complete implementations only — no stubs, no TODOs, no "will implement later". Every function must work, every import must point to a real file, every function call must reference a real function.
Also requires: research before implementing, fix all instances of a bug (not just the reported one), and security checks (nonces, sanitisation, prepared statements).
CSS/HTML/PHP/JSX visual-standards.md
WCAG 2.2 AA accessibility standards:
- Normal text: 4.5:1 contrast minimum
- Touch targets: 44x44px minimum
- Mobile-first responsive (375px → 768px → 1280px)
- Dark mode support with CSS custom properties
- Never remove focus outlines without replacing them
- Reduced motion support
PHP/WordPress wordpress.md
WordPress coding standards:
- Yoda conditions:
if ( true === $var ) - Always
$wpdb->prepare()for database queries - Nonces for all forms, capability checks for actions
- Sanitise all input, escape all output
- Proper script/style enqueueing with version numbers
- Multisite-aware table prefixes
- block.json, theme.json v3, Interactivity API, REST API patterns, SGS naming conventions
How to add a new rule
Create a .md file in C:\Users\Bean\.claude\rules\. It's picked up automatically. Add paths: frontmatter if you only want it active for certain file types.
My Agents
What agents are
Specialised sub-processes Claude can spawn to handle specific tasks. They run independently, have their own tool access, and return results to Claude. Think of them as specialist contractors — Claude is the project manager, agents are the specialists.
There are two kinds: custom agents (ones you created, stored in ~/.claude/agents/) and plugin agents (come bundled with installed plugins, auto-delegated by Claude).
Custom Agents (My Own)
Auto test-and-explain
| Model | Sonnet (smart, fast) |
|---|---|
| Can edit files? | No — reads and runs tests only |
| When it's used | After any code change, Claude automatically delegates testing here |
| What it does | Tests code, then explains results in plain English for a non-coder |
Dev wp-developer
| Model | Sonnet |
|---|---|
| Can edit files? | Yes — full read/write/edit |
| When it's used | Any WordPress task (plugins, themes, ACF, multisite, REST API, Interactivity API, SGS Framework) |
| What it does | Builds production-ready WordPress code following WP coding standards |
Dev n8n-builder
| Model | Sonnet |
|---|---|
| Can edit files? | Yes |
| When it's used | Building or debugging n8n workflows |
| What it does | Designs workflow logic, builds JSON configs, debugs executions |
Audit seo-auditor
| Model | Haiku (fast, cheap) |
|---|---|
| Can edit files? | No — report only |
| When it's used | Before launching a site, or periodic SEO checks |
| What it does | Analyses websites for SEO issues, content optimisation, and technical SEO problems |
Security ehr-security-reviewer
| Model | Sonnet |
|---|---|
| Can edit files? | No — report only |
| When it's used | Before any EHR deployment or after EHR code changes |
| What it does | HIPAA/GDPR compliance review: access controls, audit logging, encryption, data isolation |
Audit performance-auditor
| Model | Sonnet |
|---|---|
| Can edit files? | No — report only |
| When it's used | After Next.js deployment or when checking performance |
| What it does | Next.js performance: Core Web Vitals, Lighthouse, bundle size, image optimisation |
Plugin Agents (From Enabled Plugins)
Dev code-architect — feature-dev
Designs feature architectures by analysing codebase patterns. Read-only — produces blueprints, not code.
Dev code-explorer — feature-dev
Deep analysis of existing features — traces execution paths, maps architecture layers, documents dependencies. Read-only.
Audit code-reviewer — feature-dev + superpowers
Reviews code for bugs, logic errors, security vulnerabilities, and adherence to coding standards. Two versions: feature-dev (read-only report) and superpowers (reviews against your plan, has all tools).
Dev code-simplifier — code-simplifier
Simplifies and refines code for clarity, consistency, and maintainability while preserving all functionality. Can edit files.
Browser browser-user — superpowers-chrome
Analyses web content using Chrome DevTools Protocol. Read-only — inspects DOM, analyses behaviour, reads cached content.
You don't need to invoke agents manually. Claude picks the right agent based on what you're asking. But if you want to be specific, say: "Use the wp-developer agent to build this plugin."
My Commands
What commands are
Custom slash commands stored as .md files in ~/.claude/commands/. When you type /command-name, Claude loads the prompt from that file. They're saved prompts — type the shortcut instead of explaining what you want every time.
| Command | What It Does | When to Use |
|---|---|---|
/morning | Daily planning — pulls tasks, suggests priorities, reminds about physio | Start of every work session |
/brain-dump | Structures rambling thoughts into actions, ideas, decisions, worries | When you have ideas or worries to process |
/linkedin | Creates LinkedIn post in your voice with hook, story, takeaway, CTA | When you have a story or insight to share |
/deploy-check | 30+ point checklist: security, performance, SEO, accessibility, WP-specific | Before any WordPress site goes live |
/deploy-nextjs | Pre-deployment checklist for Next.js/Vercel: build, lint, meta tags, critical files | Before deploying to Vercel |
/a11y-audit | Full WCAG 2.2 AA accessibility audit on every page of a website | Before launching or after major UI changes |
/handoff | Session summary: completed, current state, blockers, next priorities | End of every work session |
How to add a new command
Create a .md file in C:\Users\Bean\.claude\commands\. The filename becomes the command name (e.g. review.md = /review).
My Skills
What skills are
Skills are like commands but come bundled with plugins. Invoked the same way — type /skill-name — but you don't manage the files yourself. Skills are more powerful than commands because they can include complex multi-step logic, tool orchestration, and specialised prompts.
Key difference: Commands = your own custom prompts. Skills = prompts that come with plugins.
Planning Development & Planning Skills
| Skill | Plugin | What It Does |
|---|---|---|
/brainstorming | superpowers | Explores requirements before building anything. Should be used before creative work |
/writing-plans | superpowers | Creates detailed implementation plans for multi-step tasks |
/systematic-debugging | superpowers | Methodical debugging — structured, not guesswork |
/test-driven-development | superpowers | Write tests before implementation code |
/feature-dev | feature-dev | Guided feature development with codebase understanding |
/subagent-driven-development | superpowers | Execute plans using parallel agents |
/dispatching-parallel-agents | superpowers | Run 2+ independent tasks in parallel |
/executing-plans | superpowers | Execute a written plan in a separate session with review checkpoints |
Quality Code Quality Skills
| Skill | Plugin | What It Does |
|---|---|---|
/requesting-code-review | superpowers | Get code reviewed after finishing a feature |
/receiving-code-review | superpowers | Process review feedback before implementing suggestions |
/verification-before-completion | superpowers | Run verification before claiming work is done |
/finishing-a-development-branch | superpowers | Guide completion of dev work — integration options |
Git Git Skills
| Skill | Plugin | What It Does |
|---|---|---|
/commit | commit-commands | Create a git commit |
/commit-push-pr | commit-commands | Full commit + push + PR workflow |
/clean_gone | commit-commands | Clean up branches deleted on remote |
/using-git-worktrees | superpowers | Create isolated git worktrees for feature work |
Content Writing & Content Skills
| Skill | Plugin | What It Does |
|---|---|---|
/writing-clearly-and-concisely | elements-of-style | Apply Strunk's rules to any prose — docs, commits, error messages |
/frontend-design | frontend-design | Create distinctive, production-grade web interfaces |
/playground | playground | Create interactive HTML playgrounds — visual configurators |
Web Web & Scraping Skills
| Skill | Plugin | What It Does |
|---|---|---|
/firecrawl-cli | firecrawl | Unified web tool: scrape, search, crawl, map, extract. Replaces WebFetch/WebSearch |
/browsing | superpowers-chrome | Direct browser control via Chrome DevTools Protocol |
Media Image & Media Skills
| Skill | Plugin | What It Does |
|---|---|---|
/generate | nano-banana-pro | Generate images from text descriptions |
Memory Memory & Search Skills
| Skill | Plugin | What It Does |
|---|---|---|
/search-conversations | episodic-memory | Search past Claude Code conversations |
/remembering-conversations | episodic-memory | Recall past decisions, solutions, lessons when stuck |
WordPress WordPress Domain Skills
These activate automatically when working on WordPress files. You can also invoke them directly.
| Skill | What It Does |
|---|---|
/wordpress-router | Classify a WP repo and route to correct workflow |
/wp-block-development | Gutenberg block development (block.json, attributes, render.php) |
/wp-block-themes | Block themes (theme.json, templates, patterns, style variations) |
/wp-interactivity-api | Interactivity API (data-wp-* directives, store, state, actions) |
/wp-performance | WordPress backend performance (profiling, queries, caching) |
/wp-plugin-development | Plugin architecture, hooks, admin UI, security, packaging |
/wp-rest-api | REST API endpoints, controllers, schema, authentication |
/wp-wpcli-and-ops | WP-CLI operations, search-replace, db management, scripting |
Design Design & UI Skills
| Skill | What It Does |
|---|---|
/tailwind-design-system | Build scalable design systems with Tailwind CSS v4 |
/ui-ux-pro-max | 50 styles, 21 palettes, 50 font pairings, 20 chart types, 9 stacks |
/vercel-react-best-practices | React/Next.js performance optimisation from Vercel Engineering |
Meta Claude Code Development Skills
For building plugins, skills, agents, and other Claude Code extensions.
| Skill | Plugin | What It Does |
|---|---|---|
/revise-claude-md | claude-md-management | Update CLAUDE.md with learnings from the session |
/claude-md-improver | claude-md-management | Audit and improve CLAUDE.md files |
/claude-automation-recommender | claude-code-setup | Analyse codebase and recommend automations |
/working-with-claude-code | superpowers-dev | Comprehensive Claude Code documentation reference |
/developing-claude-code-plugins | superpowers-dev | Patterns and workflows for building Claude Code plugins |
/using-superpowers | superpowers | How to find and use superpowers skills |
/writing-skills | superpowers | Create, edit, or verify skills |
/fix-vscode-bash-path | bean-tools (local) | Fix Git Bash path error after VS Code extension updates |
Disabled Currently Disabled Skills
These skills exist but their parent plugins are disabled. Enable the plugin in settings.json to use them.
| Skills | Plugin | Why Disabled |
|---|---|---|
| Notion skills (create-page, create-task, search, tasks:build, etc. — 10 skills) | Notion | Plugin disabled |
| Plugin development skills (create-plugin, agent-development, etc. — 7 skills) | plugin-dev | Plugin disabled |
| Hugging Face skills (CLI, datasets, training, etc. — 8 skills) | huggingface-skills | Plugin disabled |
| Agent SDK skill (new-sdk-app) | agent-sdk-dev | Plugin disabled |
My Hooks
What hooks are
Shell commands that Claude fires automatically when certain events happen. Unlike skills (which you invoke manually), hooks run on their own — before or after tool calls, when a conversation stops, or when notifications arrive.
Think of it like this: Rules tell Claude what to think. Skills tell Claude what to do when asked. Hooks tell Claude what to run automatically at specific moments.
Active Block .env edits
| Type | PreToolUse |
|---|---|
| Matcher | Write | Edit |
| File | ~/.claude/hooks/block-env.py |
| Timeout | 5 seconds |
| What it does | Prevents Claude from writing or editing any .env file. Exits with code 2 to block the operation. Manage secrets manually outside Claude Code. |
Active Auto-lint JS/TS
| Type | PostToolUse |
|---|---|
| Matcher | Write | Edit |
| File | ~/.claude/hooks/auto-lint.py |
| Timeout | 15 seconds |
| What it does | Runs npx eslint --fix on .ts, .tsx, .js, .jsx, .mjs files after every Write/Edit. Skips node_modules and .next. |
Reference Hook Types Explained
| Hook Type | When It Fires | Example Use Case |
|---|---|---|
| PreToolUse | Before Claude uses a tool (e.g. before editing a file) | Block edits to protected files, require confirmation for destructive ops |
| PostToolUse | After Claude uses a tool (e.g. after writing a file) | Auto-format code after saving, run linting, log changes |
| Stop | When Claude finishes its response | Auto-run tests after every response, update a log |
| Notification | When Claude sends a notification | Route notifications to Slack, email, or a dashboard |
Ideas Hooks Worth Adding
| Hook | Type | What It Would Do |
|---|---|---|
| PHP syntax check | PostToolUse (Write|Edit) | Run php -l after every PHP file edit — catches fatal errors instantly |
| Run tests on stop | Stop | Auto-run test suite after Claude finishes a response |
How to add a hook
Edit C:\Users\Bean\.claude\settings.json and modify the hooks key. Or enable the plugin-dev plugin and use the /hook-development skill for guided setup.
My Plugins
What plugins are
Bundles of skills, agents, hooks, and MCP servers that extend Claude Code's capabilities. Installed from marketplaces and managed in settings.json. Each plugin can contribute multiple skills and agents.
Enabled Active Plugins (22 + 1 local)
| Plugin | Source | What It Provides |
|---|---|---|
| superpowers | superpowers-marketplace | 14 skills (brainstorming, debugging, planning, code review, TDD, etc.) + code-reviewer agent |
| commit-commands | claude-plugins-official | 3 skills (commit, commit-push-pr, clean_gone) |
| episodic-memory | superpowers-marketplace | 2 skills (search-conversations, remembering-conversations) + search agent |
| elements-of-style | superpowers-marketplace | 1 skill (writing-clearly-and-concisely) |
| superpowers-chrome | superpowers-marketplace | 1 skill (browsing) + browser-user agent |
| superpowers-developing-for-claude-code | superpowers-marketplace | 2 skills (developing plugins, working with Claude Code) |
| frontend-design | claude-plugins-official | 1 skill (frontend-design) |
| nano-banana-pro | buildatscale-claude-code | 1 skill (generate images) |
| feature-dev | claude-plugins-official | 1 skill (feature-dev) + 3 agents (code-architect, code-explorer, code-reviewer) |
| code-simplifier | claude-plugins-official | 1 agent (code-simplifier) |
| claude-md-management | claude-plugins-official | 2 skills (revise-claude-md, claude-md-improver) |
| claude-code-setup | claude-plugins-official | 1 skill (claude-automation-recommender) |
| firecrawl | claude-plugins-official | Unified web tool: scrape, search, crawl, map, extract |
| playground | claude-plugins-official | 1 skill (interactive HTML playgrounds) |
| context7 | claude-plugins-official | MCP server for library documentation lookup |
| github | claude-plugins-official | MCP server for GitHub operations (issues, PRs, code search) |
| playwright | claude-plugins-official | Browser automation tools (navigate, click, fill, screenshot) |
| supabase | claude-plugins-official | Supabase integration tools |
| security-guidance | claude-plugins-official | Security best practices guidance |
| learning-output-style | claude-plugins-official | Output style — learning mode with educational insights |
| php-lsp | claude-plugins-official | PHP Language Server Protocol support |
| typescript-lsp | claude-plugins-official | TypeScript Language Server Protocol support |
| bean-tools | local | 1 skill (fix-vscode-bash-path) |
Disabled Disabled Plugins
| Plugin | Why Disabled |
|---|---|
| Notion | Plugin disabled — Notion MCP still works via Claude Desktop |
| serena | Semantic code analysis — not currently needed |
| plugin-dev | Plugin development tools — enable when building plugins |
| agent-sdk-dev | Agent SDK development — enable when building SDK apps |
| huggingface-skills | ML model tools — not currently needed |
| pyright-lsp | Python type checking — enable for Python projects |
| circleback | Not currently needed |
| linear | Not using Linear for project management |
| double-shot-latte | Not currently needed |
| superpowers (official) | Using superpowers-marketplace version instead |
| explanatory-output-style | Replaced by learning-output-style |
| gopls, csharp, rust-analyzer, jdtls, clangd, swift, kotlin, lua LSPs | Language servers for languages you don't use |
Managing plugins
Enable/disable: Edit enabledPlugins in C:\Users\Bean\.claude\settings.json — set to true or false.
Performance tip: Each enabled plugin adds to context. Disable plugins you're not using to keep Claude faster.
MCP Servers (Data Connections)
What MCP servers are
MCP (Model Context Protocol) servers connect Claude to external services — your Todoist, Notion, GitHub, etc. They run in the background and give Claude tools to interact with those services. You don't invoke them directly; Claude uses them when relevant.
Important: Claude Desktop and Claude Code have SEPARATE MCP servers. They don't share configs. A server running in Claude Desktop is NOT automatically available in Claude Code, and vice versa.
Claude Code MCP Servers
Available when working in Claude Code (VS Code).
Plugin From Plugins (always available when plugin is enabled)
| MCP Server | Plugin | What It Connects To | Example Prompt |
|---|---|---|---|
| Context7 | context7 | Library documentation | "Get WordPress REST API docs from context7" |
| GitHub | github | Your GitHub repos | "List open PRs on my repo" |
| Playwright | playwright | Browser automation | "Navigate to this URL and take a screenshot" |
| Supabase | supabase | Supabase databases | "Show my Supabase tables" |
| Firecrawl | firecrawl | Web scraping service | "Scrape this URL and summarise" |
| Episodic Memory | episodic-memory | Past conversation history | "What did I decide about the EHR auth system?" |
Global From Global Config (~/.claude/.mcp.json)
| MCP Server | What It Connects To | Example Prompt |
|---|---|---|
| a11y-accessibility | WCAG 2.2 accessibility audits via axe-core | "Test this page for accessibility" |
Remote From claude.ai (synced to Claude Code)
| MCP Server | What It Connects To | Example Prompt |
|---|---|---|
| ICD-10 Codes | Medical coding database (2026 code sets) | "Look up ICD-10 code for diabetes type 2" |
Project From Project-Level .mcp.json
| Project | Servers |
|---|---|
| booking-system | context7 (redundant with plugin), a11y-accessibility (now global) |
| small-giants-studio | a11y (uses a11y-mcp — different package from global a11y-mcp-server) |
Claude Desktop MCP Servers
Available in the Claude Desktop app only. Config: C:\Users\Bean\AppData\Roaming\Claude\claude_desktop_config.json
Active Claude Desktop Servers (14)
| MCP Server | What It Connects To | Status |
|---|---|---|
| Todoist | Your Todoist account — tasks, projects, labels | Running |
| Notion API | Your Notion workspace — pages, databases, comments | Running |
| Memory | Persistent knowledge graph — store/recall facts | Running |
| GitHub | Your GitHub repos (separate from Claude Code plugin) | Running |
| n8n | Your n8n instance at n8n.smallgiantsstudio.cloud | Running |
| Filesystem | Local files (Desktop, Downloads, .claude, Pictures, Documents, AppData) | Running |
| Fetch | HTTP requests (Python mcp_server_fetch) | Running |
| Sequential Thinking | Step-by-step reasoning for complex problems | Running |
| Chrome DevTools | Browser control via Chrome DevTools Protocol | Running |
| Playwright | Browser automation (separate from Claude Code plugin) | Running |
| Quran | Quran text lookup (custom local MCP) | Running |
| Aladhan | Prayer times API (Python aladhan_mcp.server) | Running |
| Cloudflare | Cloudflare account management | Running |
| WordPress | WordPress site management via claudeus-wp-mcp | Not yet configured |
DXT DXT Extensions (managed by Claude Desktop)
| Extension | What It Does |
|---|---|
| Windows-MCP | Windows system operations |
| Vibe Prospecting | Prospecting tools (Explorium AI) |
| Context7 | Library documentation (separate from Claude Code plugin) |
Installed but not active: gdrive, gcal, gmail, mcp-registry, Claude in Chrome
My Project Memory
What project memory is
Claude Code maintains persistent memory files per project. Each MEMORY.md stores patterns, decisions, and context that persist across conversations. Claude reads these automatically when you work in that project directory.
Memory File Locations
| Project | Path |
|---|---|
| Global (home dir) | ~\.claude\projects\C--Users-Bean\memory\MEMORY.md |
| This project (.claude) | ~\.claude\projects\c--Users-Bean--claude\memory\MEMORY.md |
| Booking System | ~\.claude\projects\c--Users-Bean-Projects-booking-system\memory\MEMORY.md |
| Indus Foods / SGS Framework | ~\.claude\projects\c--Users-Bean-Projects-indus-foods\memory\MEMORY.md |
| Small Giants Studio | ~\.claude\projects\c--Users-Bean-Projects-small-giants-studio\memory\MEMORY.md |
| Small Giants WP | ~\.claude\projects\c--Users-Bean-Projects-small-giants-wp\memory\MEMORY.md |
Session Planning Files
| File | What It Is |
|---|---|
| ~/.claude/NEXT-SESSION-PROMPT.md | Master prompt template for starting new sessions |
| ~/.claude/FUTURE-SESSION-PROMPTS.md | 5 ready-to-paste prompts for upcoming tasks |
Troubleshooting
Claude isn't following my rules
Check ~/.claude/rules/ — files must end in .md. Run /revise-claude-md to audit.
An agent isn't loading
Check ~/.claude/agents/ — file must have valid YAML frontmatter (name, description, tools, model between --- lines).
A command isn't showing up
Check ~/.claude/commands/ — file must be .md. Restart Claude Code if just created.
A skill isn't working
Check the parent plugin is enabled in settings.json under enabledPlugins. Must be true.
I forgot what tools I have
Open this guide, or type "what tools do I have?" and Claude will list them.
MCP server not connecting
Check config file for the right platform (Claude Desktop: claude_desktop_config.json, Claude Code: .mcp.json or plugin).
Plugin not working
Check settings.json enabledPlugins — must be true.
Hook not firing
Check settings.json hooks section. Matcher must match tool name. Check Python is on PATH. Check timeout isn't too short.
.env edit got blocked
That's the block-env.py hook working correctly. Manage secrets manually outside Claude Code.
ESLint auto-fix not running
The auto-lint hook only runs on .ts/.tsx/.js/.jsx/.mjs files. Check project has eslint config.
I need to add a new project
Create a CLAUDE.md file in the project root with project-specific instructions. Claude reads it automatically.
Claude is being too slow
Large plugins/MCPs use context. Disable unused plugins in settings.json. Use Haiku agents for simple tasks.
Claude forgot what we discussed
Use /search-conversations to find past discussions. Or check ~/.claude/projects/*/memory/MEMORY.md.
I want to restart fresh
Start a new conversation. CLAUDE.md and rules persist. Use /handoff first to save progress.
VS Code bash path error after update
Type /fix-vscode-bash-path.
Plugin cache growing large
The ~/.claude/plugins/cache/ directory accumulates temp files. Safe to delete temp_git_* folders.
Notion skills not working
The Notion plugin is currently disabled. Enable it in settings.json or use Notion MCP via Claude Desktop instead.
Reference
Concepts Quick Reference
| Concept | What It Is | Where It Lives | How to Use It |
|---|---|---|---|
| Rule | Permanent instruction, always active | ~/.claude/rules/*.md | Automatic — Claude reads them every conversation |
| Command | Custom slash command (your own) | ~/.claude/commands/*.md | Type /command-name |
| Skill | Slash command from a plugin | Inside plugin bundles | Type /skill-name |
| Agent | Specialist sub-process | ~/.claude/agents/*.md or plugins | Claude delegates automatically, or you say "use the X agent" |
| Hook | Auto-firing shell command | settings.json under hooks | Fires automatically on matching events |
| Plugin | Bundle of skills + agents + hooks | Installed via marketplace | Enable/disable in settings.json |
| MCP Server | Connection to external service | settings.json, .mcp.json, or plugin | Claude uses them when relevant |
| CLAUDE.md | Project-specific instructions | Project root directory | Claude reads it when you work in that directory |
| MEMORY.md | Persistent per-project memory | ~/.claude/projects/*/memory/ | Claude reads/writes automatically |
File Locations Quick Reference
| What | Where |
|---|---|
| Global instructions | C:\Users\Bean\.claude\CLAUDE.md |
| Rules (always active) | C:\Users\Bean\.claude\rules\ |
| Agents | C:\Users\Bean\.claude\agents\ |
| Commands | C:\Users\Bean\.claude\commands\ |
| Hooks | C:\Users\Bean\.claude\hooks\ |
| Claude Code settings | C:\Users\Bean\.claude\settings.json |
| Claude Code global MCP | C:\Users\Bean\.claude\.mcp.json |
| Installed plugins | C:\Users\Bean\.claude\plugins\installed_plugins.json |
| Local plugins | C:\Users\Bean\.claude\plugins\local\ |
| Project memory | C:\Users\Bean\.claude\projects\*\memory\MEMORY.md |
| Session planning | C:\Users\Bean\.claude\NEXT-SESSION-PROMPT.md |
| Future prompts | C:\Users\Bean\.claude\FUTURE-SESSION-PROMPTS.md |
| Claude Desktop config | C:\Users\Bean\AppData\Roaming\Claude\claude_desktop_config.json |
| SSH config | C:\Users\Bean\.ssh\config |
| This guide (markdown) | C:\Users\Bean\.claude\MY-SETUP-GUIDE.md |
| This guide (HTML) | C:\Users\Bean\.claude\setup-guide.html |