nhouseholder — AI Agent Skills on SkillsAuth

cleanup-old-files

Autonomous registry cleanup — detects and archives stale, outdated, or duplicate files across local, iCloud, and GitHub repos so no AI agent picks up the wrong version. NEVER deletes — always archives with clear markers. Fires automatically at session start in project repos, or on-demand when agent confusion is diagnosed.

devops1

code-reviewer

Comprehensive code review skill for TypeScript, JavaScript, Python, Swift, Kotlin, Go. Includes automated code analysis, best practice checking, security scanning, and review checklist generation. Use when reviewing pull requests, providing code feedback, identifying issues, or ensuring code quality standards.

development1

webapp-testing

Toolkit for interacting with and testing local web applications using Playwright. Supports verifying frontend functionality, debugging UI behavior, capturing browser screenshots, and viewing browser logs.

tools1

calibrated-confidence

Makes Claude honest about what it knows vs what it's guessing. Dynamically adjusts speed, depth, and communication based on confidence level. High confidence = move fast, low confidence = slow down and flag it. Prevents false certainty and false hedging. Always-on metacognitive skill.

data-ai1

context-checkpoint

Monitors context window usage, creates compact recovery checkpoints before compaction, and shows progress updates at task milestones. Prevents the

testing1

form-filler-outer

Fill out web forms, PDF forms, and applications on behalf of Nicholas using his saved personal info profile. Triggers on "fill out", "fill in", "complete this form", "application", "enrollment form", or any request to enter personal data into a form. Works with browser forms (Chrome MCP), PDF forms (PDF tools), and Word/DOCX forms.

tools1

skills/_archived/king-mode

# KING MODE — Senior Frontend Architect & Avant-Garde UI Designer **ROLE:** Senior Frontend Architect & Avant-Garde UI Designer. **EXPERIENCE:** 15+ years. Master of visual hierarchy, whitespace, and UX engineering. ## 1. OPERATIONAL DIRECTIVES (DEFAULT MODE) - **Follow Instructions:** Execute the request immediately. Do not deviate. - **Zero Fluff:** No philosophical lectures or unsolicited advice in standard mode. - **Stay Focused:** Concise answers only. No wandering. - **Output First:** P

development1

know-what-you-dont-know

Detects when Claude is about to implement domain-specific logic it hasn't verified understanding of. Forces background research BEFORE writing code in specialized domains (betting, finance, medicine, law, physics, etc.). Catches the dangerous failure mode where Claude is confident but wrong — not uncertain, but ignorant. Always-on metacognitive firewall.

development1

process-monitor

Monitor background processes (dev servers, builds, tests, watchers) during a session. Detect hung processes, runaway resource usage, port conflicts, and zombie processes. Provide status updates only when relevant. Kill unnecessary tasks when safe. Low-overhead self-awareness skill.

development1

sanity-check

Before executing a request that could make things worse, waste significant effort, or introduce problems, pause and respectfully flag the concern. Recommend a better approach if one exists. Never blindly execute a bad idea — but never be condescending about it either. Automatic skill that fires only when genuine risk is detected.

testing1

smart-model-router

Classifies each prompt as Sonnet-tier, Opus-tier, or Opus-1M-tier based on task complexity and context requirements. Injects model recommendation when current model is mismatched. Passive skill with near-zero overhead — only fires a visible recommendation when a mismatch is detected.

data-ai1

take-your-time

Match effort to prompt complexity. A 20-bullet detailed prompt deserves 20 individual implementations, not one rushed pass. Prevents AI slop by treating each requirement with the same quality as if it were its own standalone prompt. Know when to go slow and meticulous, and when to go fast. Always-on quality discipline skill.

development1

adaptive-voice

Match the user's communication energy and pace. Terse when they're in flow, detailed when they're learning, calm when they're frustrated. Always leads with the answer/action (answer first, never preamble). Read the room from message length, punctuation, word choice, and context. Always-on awareness skill with zero overhead.

development1

audience-monetization

Turn followers into paying customers. Reviews current business model and builds a monetization plan with offer ideas, pricing structure, funnel design, and the content angles that naturally move people from follower to buyer. Triggers on "monetization strategy", "how to monetize", "turn followers into customers", "offer stack", "pricing strategy", "follower to buyer".

development1

autoresearch

Systematic skill and prompt optimization using Karpathy's autoresearch method. Diagnose failure patterns, build scoring checklists, run one-change-at-a-time improvement loops, and extract reusable rules from the changelog. Use when asked to improve a skill, optimize a prompt, or make any repeatable task consistently better.

development1

codebase-cartographer

Build a mental architecture map of the codebase at session start. Know every directory's purpose, key files, data flow patterns, and entry points. Enables instant navigation and informed decisions without redundant file exploration. Automatic skill that fires at session start.

development1

competitor-gap-analysis

Find whitespace in crowded niches. Analyzes what competitors avoid, underuse, and get wrong across 5 dimensions — topic gaps, format gaps, tone gap, positioning gap, and quick wins. Produces a competitive whitespace map with clear action items. Triggers on "competitor analysis", "gap analysis", "competitive research", "whitespace map", "what are competitors missing", "niche gaps".

development1

content-calendar

Generate a full 30-day content calendar with daily content ideas, post formats, core message angles, and goals (reach, trust, conversion). Reads from brand context, audience profile, content pillars, and strategy files. Triggers on "content calendar", "30-day plan", "what should I post this month", "content schedule", "monthly content plan".

development1

content-pillars

Build 4-5 content pillars mapped to funnel stages (Awareness, Trust, Authority, Conversion). Each pillar includes example topics, platform formats, content ratios, repurposing strategy, and stage-aware lead pillar recommendation. Triggers on "content pillars", "what should I post about", "content categories", "content themes", "pillar strategy", "content pillar matrix".

development1

finishing-a-development-branch

Use when implementation is complete, all tests pass, and you need to decide how to integrate the work - guides completion of development work by presenting structured options for merge, PR, or cleanup

development1

impeccable-design

Expert design enhancement layer. Provides deep reference library (typography, color/OKLCH, motion, spatial, interaction, responsive, UX writing), AI slop detection checklist, and design normalization. Use alongside frontend-design for any site build, redesign, or visual work. Automatically loaded by site-redesign, design-options, and site-review.

development1

iterative-retrieval

Pattern for progressively refining context retrieval to solve the subagent context problem

data-ai1

mmalogic

Dedicated MMALogic/OctagonAI website agent. This command carries ALL domain knowledge for mmalogic.com — betting rules, canonical paths, anti-patterns, verification checklists, ground truth validation, and learned errors. Use this for ANY task invo

development1

pre-debug-check

Checks anti-patterns, past error solutions, and mid-execution barriers before any fix. Auto-triggers on errors, test failures, or build problems. Must fire before systematic-debugging.

development1

prompt-improver

This skill enriches vague prompts with targeted research and clarification before execution. Should be used when a prompt is determined to be vague and requires systematic research, question generation, and execution guidance.

data-ai1

refactoring-ui

Audit and fix visual design issues using the Refactoring UI system — hierarchy, spacing, shadows, color, typography. Use when the UI looks "off" but you can't explain why, when asked to fix the design, improve visual hierarchy, or refine a color palette.

development1

review-handoff

Read the most recent handoff for this project and orient yourself. The FIRST thing you do when starting a new session.

tools1

site-recover

Recover the last working version of a site after a destructive deploy, accidental overwrite, or data loss. This is the EMERGENCY command — use when production is broken and you need to get back to a known-good state.

development1

skill-awareness

MANDATORY first-thought on every prompt: scan the available skills list and use any matching skill before doing work manually. Always-on.

data-ai1

task-router

Automatically routes tasks to Opus 4.6 or Sonnet 4.6 based on complexity. Opus for debugging, planning, strategy, experimentation. Sonnet for execution, simple edits, cosmetic changes, following existing plans. Fires on every message to classify and route.

development1

ufc-hypothesis-tester

UFC algorithm experiment runner for mmalogic/OctagonAI — handles data formats, baseline workflow, activation verification, coefficient sweeps, and mandatory experiment logging. Fires on "test a theory", "hypothesis", "parameter sweep", "coefficient sweep", "would X improve picks", "test this in the backtester", or any edit to UFC_Alg_v4_fast_2026.py followed by a backtest.

development1

update-courtside

Dedicated Courtside AI website agent. This command carries ALL domain knowledge for courtside-ai.pages.dev — NBA+NCAA algorithms, betting systems, canonical paths, integrations, anti-patterns, and learned errors.

development1

update-enhancedhealth

Dedicated Enhanced Health AI website agent. This command carries ALL domain knowledge for enhancedhealthai.com — Next.js 15, TypeScript, Prisma, Cloudflare Workers, health tech domain.

development1

using-git-worktrees

Use when starting feature work that needs isolation - creates isolated git worktrees with smart directory selection and safety verification

tools1

skills/website-guardian

--- name: website-guardian description: "MANDATORY enforcement for web app changes. Snapshot baseline before edits, verify after, and run root-cause analysis with permanent logging on any bug. Fires on every file change in a webapp directory." weight: light triggers: - "update the website", "deploy", "fix the site", "the site is broken" - after ANY deploy or build (post-deploy verification) - when a user reports a visual bug or broken feature - when a screenshot shows something wrong -

development1

whats-next

Expert strategic advisor scoped to the CURRENT PROJECT ONLY. Reviews this project's handoff, memory, anti-patterns, git state, and codebase — then produces expert-level recommendations for what to do next.

development1

progress-tracker

Track and display progress across multi-step tasks. Shows completion percentage, elapsed time, task count, and milestone updates. Fires at natural task boundaries during multi-step work — not after every response. Gives the user visibility into complex work without cluttering simple tasks.

tools1

site-review

Multi-agent site review panel. Dispatches senior-dev, frontend, and backend agents in parallel to review the current project from a **functional perspective** — not just bugs, but architecture, UX, performance, and improvement opportunities.

development1

data-consistency-check

Validates that displayed data is mathematically and logically consistent before claiming correctness. Catches impossible statistics (profit with 0 wins, implausible per-unit returns, mismatched totals). Fires on any data display, dashboard, stats page, table, or summary card.

testing1

skills/profit-driven-development

--- name: profit-driven-development description: The north star for all sports prediction work — future predictive accuracy and profitability. Every backtest, every parameter change, every algorithm decision must be evaluated against one question: will this make the NEXT picks more correct and more profitable? Prevents getting lost in endless optimization loops that overfit the past. Always-on for sports code sessions. weight: passive --- # Profit-Driven Development — Never Lose Sight of the Go

development1

react-best-practices

Comprehensive React and Next.js performance optimization guide with 40+ rules for eliminating waterfalls, optimizing bundles, and improving rendering. Use when optimizing React apps, reviewing performance, or refactoring components.

development1

ui-design-system

UI design system toolkit for Senior UI Designer including design token generation, component documentation, responsive design calculations, and developer handoff tools. Use for creating design systems, maintaining visual consistency, and facilitating design-dev collaboration.

tools1

webapp-testing

Toolkit for interacting with and testing local web applications using Playwright. Supports verifying frontend functionality, debugging UI behavior, capturing browser screenshots, and viewing browser logs.

tools1

ui-ux-pro-max

UI/UX design intelligence. 50 styles, 21 palettes, 50 font pairings, 20 charts, 9 stacks (React, Next.js, Vue, Svelte, SwiftUI, React Native, Flutter, Tailwind, shadcn/ui). Actions: plan, build, create, design, implement, review, fix, improve, optimize, enhance, refactor, check UI/UX code. Projects: website, landing page, dashboard, admin panel, e-commerce, SaaS, portfolio, blog, mobile app, .html, .tsx, .vue, .svelte. Elements: button, modal, navbar, sidebar, card, table, form, chart. Styles: glassmorphism, claymorphism, minimalism, brutalism, neumorphism, bento grid, dark mode, responsive, skeuomorphism, flat design. Topics: color palette, accessibility, animation, layout, typography, font pairing, spacing, hover, shadow, gradient. Integrations: shadcn/ui MCP for component search and examples.

tools1

site-redesign

Redesign a website/webapp with production-grade quality using an agent pipeline architecture. Each phase dispatches a specialized agent. Agents communicate through shared memory files in `_redesign/`. Context is handed off explicitly between phases.

development1

security-review

Use this skill when adding authentication, handling user input, working with secrets, creating API endpoints, or implementing payment/sensitive features. Provides comprehensive security checklist and patterns.

development1

compactor

Unified context management and session continuity skill. Combines total-recall, strategic-compact, /ledger, and session continuity. Runs in background to preserve critical context across compaction and sessions.

tools1

prompt-architect

Restructures prompts for optimal execution, detects intent to route to skills/commands, and stays anchored to the original objective during long sessions. Combines prompt processing, intent detection, and drift prevention. Always-on.

data-ai1

audience-psychology

Deep audience psychology breakdown + ICP builder using Jobs-to-be-Done framework. Maps frustrations, desires, fears, daily habits, and content consumption patterns into specific messaging angles and content topics. Builds Ideal Customer Profiles with functional/emotional/social jobs. Triggers on "audience research", "who is my audience", "audience psychology", "target audience breakdown", "messaging angles", "ICP", "ideal customer profile", "jobs to be done".

development1

update-researcharia

Dedicated Research Aria website agent. This command carries ALL domain knowledge for researcharia.com — Workers Sites, D1 database, paper research features, and learned errors from the CATASTROPHIC 2026-03-26 incident.

development1

canvas-design

Create beautiful visual art in .png and .pdf documents using design philosophy. You should use this skill when the user asks to create a poster, piece of art, design, or other static piece. Create original visual designs, never copying existing artists' work to avoid copyright violations.

documentation1

audit

Audit codebases for security issues, code quality problems, and hardcoded secrets. Scans Python and JavaScript files for API keys, credentials, and anti-patterns. Fixes issues and commits with structured messages. Use when asked to audit, scan for secrets, check security, or review code quality.

development1

audit-context-building

Enables ultra-granular, line-by-line code analysis to build deep architectural context before vulnerability or bug finding.

development1

backtest

Run backtests for prediction models (UFC, sports betting). Ensures visible output via tee, compares against baseline accuracy, and commits improvements with structured messages. Enforces walk-forward integrity, overfitting awareness, and future predictive accuracy as the

testing1

skills/backtestor-quality-control

--- name: backtestor-quality-control description: "Quality control audit for backtestors across sports projects — enforces walk-forward integrity, real Vegas odds, correct unit math, complete bet type coverage, data caching, and single canonical backtestor per sport. Fires when building, modifying, auditing, or running any backtestor." weight: light triggers: - building or modifying any backtestor - "audit the backtestor", "check the backtestor" - running any backtest - creating a new sp

development1

brainstorming

Explore intent, requirements, and design before implementing complex features. Triggers on 3+ file changes or ambiguous multi-approach tasks. Skips automatically for clear-scope work.

data-ai1

canvas-design

Create beautiful visual art in .png and .pdf documents using design philosophy. You should use this skill when the user asks to create a poster, piece of art, design, or other static piece. Create original visual designs, never copying existing artists' work to avoid copyright violations.

documentation1

claude-api

Build apps with the Claude API or Anthropic SDK. TRIGGER when: code imports `anthropic`/`@anthropic-ai/sdk`/`claude_agent_sdk`, or user asks to use Claude API, Anthropic SDKs, or Agent SDK. DO NOT TRIGGER when: code imports `openai`/other AI SDK, general programming, or ML/data-science tasks.

development1

cleanup-old-files

Autonomous registry cleanup — detects and archives stale, outdated, or duplicate files across local, iCloud, and GitHub repos so no AI agent picks up the wrong version. NEVER deletes — always archives with clear markers. Fires automatically at session start in project repos, or on-demand when agent confusion is diagnosed.

devops1

coding-standards

Universal coding standards, best practices, and patterns for TypeScript, JavaScript, React, and Node.js development.

development1

content-research-writer

Research + writing assistant — searches the web, gathers sources, builds outlines, and drafts cited content. For blog posts, articles, case studies, or newsletters. Triggers on 'write an article', 'draft a blog post', 'research and write'.

development1

continuous-learning-v2

Instinct-based learning system that observes sessions via hooks, creates atomic instincts with confidence scoring, and evolves them into skills/commands/agents. v2.1 adds project-scoped instincts to prevent cross-project contamination.

data-ai1

deploy

Deploy projects to Cloudflare Pages/Workers with pre-flight checks, smoke tests, and automatic rollback on failure. Handles linting, testing, building, deploying, verifying, and tagging releases. Use when asked to deploy, ship, release, push to production, or go live.

development1

design-sprint

Run a structured 5-phase design sprint — problem framing, sketching, deciding, prototyping, testing. Use at the START of a project or feature, not the end. Triggers on "run a design sprint", "ideation workshop", "validate this idea", "new product", or any greenfield project kickoff.

testing1

doc-coauthoring

Structured workflow for co-authoring documentation, proposals, technical specs, or decision docs. Triggers when user mentions writing docs, drafting proposals, or creating specs.

testing1

executing-plans

Use when you have a written implementation plan to execute in a separate session with review checkpoints

testing1

expert-lens

Adopt domain-expert perspectives automatically. When the user says "you are an expert in X" or the task implies a specific professional domain, activate that expert's mental models, vocabulary, frameworks, and quality standards. Knows what real experts know — including what amateurs get wrong. Always-on detection with zero overhead when not triggered.

development1

fpf-hypotheses

Execute complete FPF cycle from hypothesis generation to decision

testing1

frontend-design

Create distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, pages, artifacts, posters, or applications (examples include websites, landing pages, dashboards, React components, HTML/CSS layouts, or when styling/beautifying any web UI). Generates creative, polished code and UI design that avoids generic AI aesthetics.

development1

frontend-design

Create distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, pages, artifacts, posters, or applications (examples include websites, landing pages, dashboards, React components, HTML/CSS layouts, or when styling/beautifying any web UI). Generates creative, polished code and UI design that avoids generic AI aesthetics.

development1

full-handoff

Generate a lean handoff document and store it in this project's repo. The LAST thing you do before ending a session.

documentation1

git-sorcery

Advanced git intelligence beyond basic add/commit/push. Smart commit messages from diffs, conflict resolution strategies, interactive history analysis, bisect for bug hunting, cherry-pick workflows, branch management, and stash operations. Automatic skill that enhances all git operations.

research1

hooked-ux

Analyze product retention using the Hook Model (trigger, action, reward, investment). Use when users aren't coming back, engagement is low, you need to improve retention, or designing a habit loop for a product feature.

data-ai1

isolate-before-iterate

Before debugging via full pipelines (backtests, builds, deploys), isolate the suspect logic in a minimal standalone test. Prevents the anti-pattern of 30+ minute feedback loops when a 5-line script would answer the question in seconds.

development1

mcp-builder

Guide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK).

tools1

mid-task-triage

When the user sends a new message while Claude is mid-task, instantly classify it as an addendum (merge into current work), a course correction (steer current task), or a queue item (address after current task completes). Never lose progress. Never stop working to ask what they meant. Always-on awareness skill with zero overhead.

tools1

opportunistic-improvement

Two modes — (1) While working, notice and fix code smells in files you're already touching. (2) When idle, suggest the highest-impact improvements across the project. Always-on mini-audit while working, proactive suggestions when idle.

development1

pattern-propagation

When a pattern is changed in one place, automatically find and update all instances across the codebase. Covers renames, API changes, style updates, config changes, and structural refactors. Automatic skill that fires when modifying shared patterns.

development1

predictive-next

After completing a task, anticipate what the user likely needs next and offer it proactively. Uses pattern recognition from the task type, codebase state, and common workflows. Automatic skill that fires after task completion — one-line suggestion, zero pressure.

development1

process-monitor

Monitor background processes (dev servers, builds, tests, watchers) during a session. Detect hung processes, runaway resource usage, port conflicts, and zombie processes. Provide status updates only when relevant. Kill unnecessary tasks when safe. Low-overhead self-awareness skill.

development1

progress-tracker

Track and display progress across multi-step tasks. Shows completion percentage, elapsed time, task count, and milestone updates. Fires at natural task boundaries during multi-step work — not after every response. Gives the user visibility into complex work without cluttering simple tasks.

tools1

qa-gate

Mandatory QA gate before delivering any feature, fix, or component. Dispatches an independent testing agent to exercise the implementation end-to-end before returning results to the user.

testing1

skills/reflexion

--- name: reflexion description: Self-refinement and comprehensive review framework. Two modes — /reflexion:reflect for quick self-assessment of output quality (checklist + confidence score), and /reflexion:critique for deep multi-dimensional review (requirements + design + code quality via reviewer agent). Opt-in only — fires on explicit user request, high-risk changes, or after 2+ failed bug fixes. weight: heavy argument-hint: "reflect" for self-assessment, "critique" for deep review, or speci

development1

receiving-code-review

Use when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementation

development1

response-recap

Provides a plain English recap ONLY after complex, multi-step work — not after every response. Fires after multi-file changes, debugging sessions, architecture decisions, or multi-step implementations. Skips for single-file edits, quick fixes, Q&A, and routine tasks. Hybrid approach that gives summaries when they matter without bloating simple responses.

development1

sanity-check

Before executing a request that could make things worse, waste significant effort, or introduce problems, pause and respectfully flag the concern. Recommend a better approach if one exists. Never blindly execute a bad idea — but never be condescending about it either. Automatic skill that fires only when genuine risk is detected.

testing1

site-audit

Run a comprehensive website/webapp audit. Finds every issue across frontend, backend, design, and functionality — then fixes the critical ones.

development1

site-debug

Debug a website/webapp systematically. Find the bug, fix it, verify the fix, log it permanently.

development1

search-first

Research-before-coding workflow. Search for existing tools, libraries, and patterns before writing custom code. Invokes the researcher agent.

tools1

security-scan

Scan your Claude Code configuration (.claude/ directory) for security vulnerabilities, misconfigurations, and injection risks using AgentShield. Checks CLAUDE.md, settings.json, MCP servers, hooks, and agent definitions.

tools1

senior-backend

Backend audit and development skill. When invoked, performs a systematic audit of ALL backend code — API routes, middleware, auth, database, storage, error handling. Finds security vulnerabilities, performance issues, and architecture problems. Also guides backend feature development with production-grade patterns. Use when asked to audit backend, review APIs, fix server issues, or build backend features.

development1

senior-frontend

Comprehensive frontend development skill for building modern, performant web applications using ReactJS, NextJS, TypeScript, Tailwind CSS. Includes component scaffolding, performance optimization, bundle analysis, and UI best practices. Use when developing frontend features, optimizing performance, implementing UI/UX designs, managing state, or reviewing frontend code.

development1

senior-prompt-engineer

World-class prompt engineering skill for LLM optimization, prompt patterns, structured outputs, and AI product development. Expertise in Claude, GPT-4, prompt design patterns, few-shot learning, chain-of-thought, and AI evaluation. Includes RAG optimization, agent design, and LLM system architecture. Use when building AI products, optimizing LLM performance, designing agentic systems, or implementing advanced prompting techniques.

development1

site-update-protocol

UFC MMA Logic website frontend update and maintenance protocol. Fires after ANY algorithm update, backtest re-run, new event scoring, or website deploy. Ensures all 5 bet types display correctly, charts render properly, tables are complete, and data is consistent across all pages. Prevents the recurring bugs of missing bet types, broken charts, and incomplete tables.

development1

skill-creator

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.

testing1

skill-creator

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.

testing1

skill-manager

Prevents skill overload. Weight class limits, conflict resolution, stack cap at 75. Always-on meta-skill.

tools1

spec-interview

Interview the user about a feature before writing any code. Produces a reusable SPEC.md artifact. Use when starting any non-trivial feature.

development1

systematic-debugging

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes

development1

strategic-compact

Suggests manual context compaction at logical intervals to preserve context through task phases rather than arbitrary auto-compaction.

tools1

skills/structured-build

--- name: structured-build description: Mandatory pipeline for new features and complex tasks. Enforces Research → Strategize → Plan → Execute → Self-Check → Verify → Deploy with gates between each phase. Prevents the #1 failure mode: jumping to code without understanding the problem. weight: light triggers: - new feature requests - complex multi-file tasks - tasks touching 3+ files or 2+ subsystems - "build me", "add feature", "implement", "new page", "new system" - routed via improve

development1

subagent-driven-development

Use when executing implementation plans with independent tasks in the current session

development1

the-fool

Use when challenging ideas, plans, decisions, or proposals using structured critical reasoning. Invoke to play devil's advocate, run a pre-mortem, red team, or audit evidence and assumptions.

development1

test-driven-development

Use when implementing any feature or bugfix, before writing implementation code

development1

token-economics

Optimizes token spend across model tiers, effort levels, and plan/execute splits. Recommends Sonnet for simple tasks, promotes Opus-plan/Sonnet-execute workflow for complex work, and calibrates effort to task complexity. Always-on awareness that shapes every response.

data-ai1

total-recall

Cross-session memory, crash-safe checkpointing, automatic handoff before context degrades, and instant resume. Handles all persistence across sessions. Always-on.

testing1

ui-design-system

UI design system toolkit for Senior UI Designer including design token generation, component documentation, responsive design calculations, and developer handoff tools. Use for creating design systems, maintaining visual consistency, and facilitating design-dev collaboration.

tools1

update-diamond

Dedicated Diamond Predictions website agent. This command carries ALL domain knowledge for diamondpredictions.com — MLB+NHL algorithms, betting systems, canonical paths, integrations, anti-patterns, and learned errors.

development1

update-mystrainai

Dedicated MyStrainAI website agent. This command carries ALL domain knowledge for mystrainai.com — cannabis strain recommendations, Cannalchemy backend, React/Vite frontend, Supabase, and learned errors.

development1

website-design-agent

Unified website design agent — creative direction, design references, critique, visual audit, and design tokens in one skill. Use for any web design task: build, redesign, review, critique, or audit.

development1

user-rules

Captures, persists, and enforces user-defined rules across sessions. When the user sets a hard constraint ("max 70 events", "always use approach X", "never do Y"), it's saved to a rules file and checked before every relevant action. Rules survive compaction, crashes, and session boundaries. Always-on enforcement skill.

testing1

using-ultraplan

Suggest /ultraplan for complex planning tasks on Claude Code CLI (2.1.91+ only). Research preview.

tools1

verification-before-completion

Use when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always

data-ai1

version-bump

Automated semantic versioning for commits. Determines major/minor/patch from the nature of changes, bumps package.json version, and formats commit messages with version prefix. Use when committing changes that warrant a version bump, or when the user asks to bump/tag/release a version.

development1

z-ai-switch

Dynamic model routing — Haiku 4.5 in picker = GLM-5 via Z AI. Opus/Sonnet = Anthropic. No restart needed. Type /z for help.

data-ai1

react-best-practices

Comprehensive React and Next.js performance optimization guide with 40+ rules for eliminating waterfalls, optimizing bundles, and improving rendering. Use when optimizing React apps, reviewing performance, or refactoring components.

development1

dispatching-parallel-agents

Parallel agent orchestrator with strict safety controls. Max 2 agents for research, max 3 for independent code tasks. Agents share context via briefings that include memory files, anti-patterns, and verified constants. NEVER parallelize data analysis. Sequential-first mindset — only parallelize when genuinely independent AND the overhead is justified.

development1

mid-task-triage

When the user sends a new message while Claude is mid-task, instantly classify it as an addendum (merge into current work), a course correction (steer current task), or a queue item (address after current task completes). Never lose progress. Never stop working to ask what they meant. Always-on awareness skill with zero overhead.

tools1

skills/_archived/profit-driven-development

--- name: profit-driven-development description: The north star for all sports prediction work — future predictive accuracy and profitability. Every backtest, every parameter change, every algorithm decision must be evaluated against one question: will this make the NEXT picks more correct and more profitable? Prevents getting lost in endless optimization loops that overfit the past. Always-on for sports code sessions. weight: light --- # Profit-Driven Development — Never Lose Sight of the Goal

development1

screenshot-dissector

Maximizes Claude's visual bug analysis when screenshots are provided during troubleshooting. Enforces a methodical, pixel-level examination instead of a quick glance — catches layout bugs, state issues, console errors, and UI regressions that surface only visually.

research1

seo-aeo

SEO and Answer Engine Optimization for portfolio sites. Covers metadata, Open Graph, JSON-LD structured data, sitemaps, robots.txt, EEAT signals, and AI-answer readiness. Use during site builds, redesigns, or audits.

development1

skills/_archived/sports-betting-intelligence

--- name: sports-betting-intelligence description: Professional sports betting domain expert — edge identification, system brainstorming, contrarian sharp logic, CLV tracking, and sport-specific knowledge for NFL, NBA, MLB, NHL, UFC, soccer, tennis, and college sports. Fires when working on prediction algorithms, discussing betting strategy, brainstorming new angles, or evaluating system profitability. The betting brain that complements backtest and profit-driven-development skills. weight: ligh

development1

type-aware-rigor

Match brainstorming depth, QA tier, and planning granularity to the TYPE of work. Site updates get lite treatment, new builds and algorithms get deep treatment. Influences brainstorming, qa-gate, writing-plans, and deploy decisions. Always-on enhancement layer.

development1

authority-positioning

Build a personal branding and authority positioning strategy. Defines what makes you THE go-to name in your niche, your unique mechanism, proof stack, and content angles that establish expertise. Triggers on "authority positioning", "personal brand", "positioning strategy", "how do I stand out", "brand differentiation".

development1

confusion-prevention

Detect when Claude is confused and force a re-orientation instead of spiraling. Prevents the "wait... actually... let me check..." pattern. Snapshot critical state before destructive actions, recognize confusion signals, and stop to re-orient. Always-on awareness skill.

testing1

content-performance-audit

Monthly social media performance audit. Tracks the 5 metrics that predict growth and sales (not vanity metrics). Diagnoses why posts flopped, finds top content angles to double down on, and delivers the ONE change to make next month. Produces a reusable monthly review template. Triggers on "monthly audit", "content performance", "what's working", "review my content", "performance review", "monthly review", "content audit".

testing1

design-critique

Structured UX critique using Nielsen's 10 Heuristics (/40 score), cognitive load assessment (8-item checklist), and persona-based testing (5 archetypes). Use during site-review, site-redesign Phase 7, or any design evaluation. Produces quantitative scores, not just opinions.

testing1

fix-loop

Self-healing CI loop — run test suite, diagnose failures, fix source code, re-run until all tests pass, then commit. Operates autonomously in a test-driven fix cycle. Use when asked to fix all tests, make tests pass, run a fix loop, or do self-healing CI.

development1

implicit-preferences

Detect patterns in user corrections and adapt without being told. If the user corrects the same type of thing repeatedly, treat it as a permanent preference for this session. Learns from behavior, not just explicit rules.

testing1

never-give-up

Never abandon a proven-valuable idea because integration failed. But also never burn tokens on endless retries. Uses an evidence-based triage to distinguish proven ideas worth fighting for from genuinely bad ones. Always-on mindset skill with built-in token discipline.

testing1

parallel-sweep

Run parallel parameter sweeps using multiple Claude Code headless agents. Each agent explores a different parameter subspace, writes results to shared SQLite database, and a summary report is generated. Use for coefficient searches, hyperparameter tuning, or any parallelizable parameter optimization.

development1

precision-reading

Read files surgically — grep first to find relevant sections, then read only those lines with offset+limit. Prevents loading entire 2000-line files when only 5-20 lines matter. Always-on efficiency skill that reduces token consumption without losing any information. Makes file reading targeted rather than exhaustive.

development1

requesting-code-review

Use when completing tasks, implementing major features, or before merging to verify work meets requirements

development1

scroll-stopper

Write high-engagement social media posts that stop the scroll. Takes a topic, writes a hook, delivers insight, and closes with a CTA that drives comments, saves, or clicks. Platform-aware formatting. Triggers on "write a post", "social media post", "scroll stopper", "write me a tweet", "Instagram caption", "LinkedIn post", "draft a post about".

tools1

social-media-strategy

Build a complete social media strategy OR audit an existing one. Covers brand positioning, content direction, audience targeting, platform strategy, monetization path, and 90-day roadmap. Two modes — BUILD (new presence) and AUDIT (diagnose + fix existing). Triggers on "social media strategy", "social strategy", "strategy audit", "audit my social", "build my brand strategy", "how should I position on social".

development1

think-efficiently

Prevents wasted tokens — pointless actions, analysis paralysis, redundant work, and disproportionate effort. Always-on efficiency filter.

testing1

update-nestwisehq

Dedicated NestWise HQ website agent. This command carries ALL domain knowledge for nestwisehq.com — Next.js, TypeScript, Prisma, Cloudflare Workers (OpenNext), family finance domain.

development1

zero-iteration

Mentally execute code before writing it. Trace inputs through logic, predict outputs, catch bugs before they exist. Reduces edit-test-fix cycles by getting it right the first time. Always-on awareness skill that activates during any code generation.

development1

ui-ux-pro-max

UI/UX design intelligence. 50 styles, 21 palettes, 50 font pairings, 20 charts, 9 stacks (React, Next.js, Vue, Svelte, SwiftUI, React Native, Flutter, Tailwind, shadcn/ui). Actions: plan, build, create, design, implement, review, fix, improve, optimize, enhance, refactor, check UI/UX code. Projects: website, landing page, dashboard, admin panel, e-commerce, SaaS, portfolio, blog, mobile app, .html, .tsx, .vue, .svelte. Elements: button, modal, navbar, sidebar, card, table, form, chart. Styles: glassmorphism, claymorphism, minimalism, brutalism, neumorphism, bento grid, dark mode, responsive, skeuomorphism, flat design. Topics: color palette, accessibility, animation, layout, typography, font pairing, spacing, hover, shadow, gradient. Integrations: shadcn/ui MCP for component search and examples.

tools1

screenshot-dissector

Maximizes Claude's visual bug analysis when screenshots are provided during troubleshooting. Enforces a methodical, pixel-level examination instead of a quick glance — catches layout bugs, state issues, console errors, and UI regressions that surface only visually.

research1

research-vault

Capture raw research (screenshots, articles, tips) and compile into a searchable wiki. Activates when user shares external content for evaluation or asks to review accumulated research.

development1

dispatching-parallel-agents

Parallel agent orchestrator with strict safety controls. Max 2 agents for research, max 3 for independent code tasks. Agents share context via briefings that include memory files, anti-patterns, and verified constants. NEVER parallelize data analysis. Sequential-first mindset — only parallelize when genuinely independent AND the overhead is justified.

development1

brainstorming

You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.

development1

deep-research

Self-directed literature review before implementation. Researches authoritative sources (papers, official docs, expert blogs) when encountering unfamiliar concepts, algorithms, APIs, or statistical methods. Triggers on 'research this', 'learn about this first', or unfamiliar domain knowledge.

development1

error-memory

Captures failed approaches, working solutions, and recurring bugs after debugging. Analyzes why reasoning went wrong (not just what broke). Escalates repeat offenders to root-cause analysis.

development1

senior-backend

Backend audit, architecture, and development — APIs, middleware, auth, DB, error handling, service boundaries, scalability, tech stack. Use for backend audits, API/architecture review, system design, or backend features.

development1

smart-clarify

When intent is ambiguous, asks a structured multiple-choice clarification (2-4 options + escape hatch) instead of guessing. Always-on — fires on ambiguity.

tools1

skills/sports-betting-intelligence

--- name: sports-betting-intelligence description: "Pro sports betting domain expert — edges, contrarian sharp logic, CLV tracking, and sport-specific knowledge (NFL, NBA, MLB, NHL, UFC, soccer, tennis, CBB). Fires on prediction algorithm work, betting strategy, brainstorming, or profitability evaluation." weight: light triggers: - working in sports prediction codebases - "brainstorm edges", "find angles", "what systems work" - "sharp money", "contrarian", "CLV", "closing line" - "is thi

development1

writing-plans

Use when you have a spec or requirements for a multi-step task, before touching code

development1

anti-slop

Zero tolerance for placeholder data in structured output. Every field must be real data or explicitly flagged as unfound. Always-on quality gate.

testing1

senior-architect

Architecture audit and system design skill. When invoked, evaluates the overall architecture — service boundaries, data flow, dependency graph, tech stack choices, scalability risks. Also guides system design for new features. Use when asked to review architecture, design a system, evaluate tech stack, or plan infrastructure.

testing1

senior-dev-mindset

Think like a senior full-stack developer who ships complete, production-ready features. Covers frontend (React, Next.js, TypeScript, Tailwind), backend (Node, Express, Python, Postgres, GraphQL), and architecture (system design, tech decisions, scalability). Infer unstated requirements from real-world context. Never leave stubs, placeholders, or TODO comments. Build the full thing every time.

development1

site-update

Update a website/webapp safely. Baseline → changes → verify → deploy. The guardrail command — ensures updates don't break what already works.

development1