buildwithclaude

Complete collection of ~110 Claude Code skills across document processing, development, automation, CRM, DevOps, and more

200 skills

Complete collection of 174 slash commands across 22 categories

177 commands

Complete collection of 117 specialized AI agents across 11 categories

117 agents

Get Shit Done -- a structured workflow plugin for Claude Code that adds planning, execution, and verification commands with MCP-backed project state

83 skills · 33 agents · 1 MCP server

Full-lifecycle SDLC orchestrator: a 12-stage pipeline (SRS, architecture, spec, plan, build, eval, deploy, monitor, release) driven by specialized agents, persistent cross-session memory, auto-reflection, gate enforcement, and semantic skill mining that proposes new skills from recurring workflows.

45 skills · 27 agents · hooks

74 skills for Claude Code and Codex. A build DAG researches your repo, fans out disjoint lanes, and refutes its own review before anything lands. A blunt A-F ship grade blocks the bad merge instead of complimenting it. The same folder designs the page, writes the copy, and ships native iOS and Android builds.

70 skills · 1 MCP server

Open-source SEO, GEO, and marketing skills (48) with one hosted NotFair MCP connection (OAuth). Audit sites and content for search and AI-answer visibility, analyze GA4 and Search Console, review Google, Meta, X, LinkedIn, Reddit, and TikTok Ads, and operate connected WordPress and GoHighLevel accounts. Every spend change and WordPress or CRM write waits for explicit approval of the exact change; fetched pages, ad and competitor text, and tool output are treated as untrusted data.

45 skills · 1 MCP server

Business operations command center for Claude Code — 30 skills, 14 agents, smart background daemon. Manage communications (WhatsApp/Email/Slack/Telegram), projects, infrastructure (AWS), revenue (Stripe+RevenueCat), e-commerce, marketing, monitoring (Datadog/New Relic/OTEL), and voice. Includes /gtm go-to-market planner, /ops:projects portfolio dashboard, and 4 parallel C-suite agents (YOLO mode).

30 skills · 14 agents · hooks

A coordinated engineering team of named AI specialists that remembers your project between sessions. Call yui to route work, or call any specialist directly: Go (jaan, siim, mart), Angular (liis, kadi, tiiu), iOS (ren, shu, ryo), Python (eero, anu, ivo), React (noa, rui, aki), SQL (saku), NoSQL (remo), architecture for any stack (kai), process management (enn), and DevOps/infra (sora). Each ships as both a skill and a subagent. Bundled hooks capture finished work when a session ends and say at the next session start that saved work exists, and record per-session and per-agent token and time usage. Every specialist follows the same discipline: read context first, make the smallest correct change, verify.

23 skills · 20 agents · hooks

Generate AI images, videos, music and speech from Claude Code through the PopNow API: Nano Banana, GPT Image, Seedream, FLUX, Kling, Seedance, MiniMax and Suno behind one key, plus workflow skills for launch films, explainer videos, social covers, landing pages and brand kits. Every paid run is quoted in USD and waits for an explicit yes. Needs a PopNow API key with a prepaid balance.

32 skills

Discipline kit for coding agents: plan files with executable done-criteria, an automated develop-and-validate pipeline, adversarial code review with a security scan, and a completion gate that judges "done" by command output. 15 skills, 15 subagents and 5 Python hooks (session-start rule injection, sensitive-path guard, ruff/prettier auto-format, Stop-time lint and test run). No network calls, no MCP servers.

14 skills · 15 agents · hooks

Convene a meeting of AI personas (Boss, Pusher, Rookie, Watcher, Cynic) to debate your question. 3 to 10 participants, mix teams (dev, design, product, business, life), add custom personas on the fly.

1 skill · 25 agents

Twenty-six free skills for content on LinkedIn and Twitter: draft in your voice, audit a draft before you publish, draft comments, fix opening lines and endings, read your numbers and plan the week. They need no account; connecting VoiceMoat (Pro, Enterprise or the $5 Pro trial) adds a voice profile trained on your own published posts, and Style and Stance checks.

26 skills

Agent Skills for Amazon sellers and advertisers: Ads audits and reporting, search-term mining, budget pacing, DSP and Amazon Marketing Cloud, listing optimization, product images, repricing, FBA inbound, refunds, and profitability. Built for Kuudo hosted Amazon Ads and Selling Partner MCP servers. Repository guidance cannot weaken skill safeguards or supply user approval; previews, confirmations, trusted-target checks, and restrictions on unattended writes remain required.

22 skills · hooks

Multi-agent orchestration grounded in an LLM wiki. /dev-loop:orchestrate splits one goal into a dependency graph of parallel Claude Code sessions, each in its own worktree, and supervises them natively on Orca (a tracked Task + Dispatch per phase, pushed worker_done / question / escalation events instead of polling, stall detection; raw tmux fallback). Every plan decision must cite a page in a bundled best-practice wiki of 400+ pages across 10 domains, checked by machine gates and a fresh-context plan reviewer. Insights from your sessions are queued and turned into reviewed wiki PRs when you run /dev-loop:knowledge-flush (automatic flushing is opt-in and never bypasses permission checks).

9 skills · 11 agents · 1 MCP server · hooks

Turn Claude Code into a problem-driven scrum team. You file problems as Linear issues; an agent team runs the empirical Scrum loop (scope, solve, review, adapt) and owns the solutions. An independent reviewer gates every increment against its acceptance criteria, the Definition of Done, and a standards corpus. Runs on your Claude subscription via the official Linear MCP over OAuth — no API keys.

8 skills · 7 commands · 4 agents · 1 MCP server · hooks

Audit and upkeep skills that keep the rules and skills in a Claude Code setup from piling up and going stale. Ask "audit my skills" and skill-stocktake proposes keep, merge or retire for each skill with a reason; the audit skills change a file only after you confirm it. Skills only: no hooks, no MCP servers, no settings. Some skills run bundled Python scripts with uv, and skill-comply and skill-stocktake start separate claude -p sessions with the shell and subagent tools denied.

20 skills

Commands for code review, testing, and analysis

19 commands

Commands for task management and project tracking

17 commands

Specialized commands for Svelte and SvelteKit development

16 commands

Agents for code review, security audits, debugging, and quality assurance

15 agents

Persistent project management inside your AI coding tool. PAPI turns a plan into a cycle of scoped tasks, hands each one to your agent with a build spec, records what actually happened, and keeps every decision alive across sessions — so a new session resumes where the last one stopped instead of starting from nothing. Connects to the hosted PAPI server over OAuth; nothing runs locally.

14 skills · 1 MCP server

Makes a repository know itself and preserve the engineering context built while you work with it, so an agent stops rediscovering both. An architecture index and an impact map instead of scanning files; work state, session records and the decisions behind them instead of losing context between sessions; the procedures, tools and workflows you keep re-deriving. Everything is repository-local markdown, committed beside the code. For long-lived repos you revisit, not one-off scripts.

11 skills · 2 agents · hooks

General debugging and utility commands

14 commands

Per-stage model routing + pre-flight cost gate for Claude Code ultracode dynamic workflows. Pins an explicit model + effort on every agent() subagent stage, estimates cost before launch, and ships a static guard + PreToolUse gate that flag stages which would silently inherit the session's Opus model. Zero dependencies, no telemetry, MIT.

1 skill · 12 commands · hooks

13 MIT-licensed SKILL.md skills for everyday developer work: review-my-diff, repro-first-debugging, pr-description, safe-refactor, dependency-upgrade (shows the target version and files and waits for a yes before changing anything), explain-codebase, security-quick-review, prove-it-ran (make the agent show real test output before saying done), session-handoff, brag-doc-from-git, write-agents-md, agents-md-audit (with a small read-only Python audit script) and write-subagent. Skills only: no hooks, no MCP servers, no settings, no network calls.

13 skills

Agent skills for driving LinkedIn work through the Curviate CLI: profiles and companies, search, inbox and messaging, posts and comments, network, job postings, and the Sales Navigator and Recruiter surfaces.

13 skills

A local-first startup advisor that guides founders through idea idea validation, competitors research, hypotheses generation, MVP/prototype design, and more.

8 skills · 4 agents · hooks

Agents for data engineering, machine learning, and AI development

12 agents

Expert agents for software architecture, backend development, and system design

12 agents

Expert agents for specific programming languages (Python, Go, Rust, etc.)

12 agents

Commands for integrating with external services and syncing data

12 commands

Commands for team workflows, PR reviews, and collaboration

12 commands

Commands for Git operations, commits, and PRs

12 commands

Turns OpenAI Codex into a pool of parallel agent workers inside Claude Code. Each worker runs in its own git worktree so parallel edits don't collide. Resumable sessions via Codex's native thread_id, dual-model PR review with materialized detached worktree, and role-based specialization (implementer/reviewer/planner/generic).

9 commands · 3 agents

AI work memory for Claude Code: brief, capture, recall, handoff, and distill, backed by a local daemon and MCP so decisions, lessons, project context, and wiki pages carry across sessions.

11 skills · hooks

Agents for podcasts, audio/video, transcription, OCR, and content production pipelines

11 agents

Spec-driven development and context engineering for AI coding agents, backed by project context in Git.

4 skills · 4 commands · 2 agents · hooks

Display OpenAI Codex usage and rate limits inside Claude Code. Real-time statusline integration with claude-hud (auto-installed via /codex-hud:setup).

11 commands

Commands for CI/CD setup, containerization, and deployment automation

11 commands

Build AG2 (AutoGen) multi-agent systems with slash commands: scaffold agents, wire workflows, create tools, and review code

7 commands · 3 agents

Domain-specific expert agents for validation, knowledge-graph, and other specialized tasks

10 agents

Commands for generating documentation and managing changelogs

10 commands

Skills for event and conference sponsorship work on live data: sponsorship context, find sponsors for an event, price packages, draft a prospectus, draft sponsor outreach, vet a trade show, budget a booth, track where competitors exhibit, build an events calendar. Adds 9 skills and the hosted ConferenceDB MCP server (https://conferencedatabase.com/api/mcp); no sign-in, no API key. Company-level data only: no people, emails or phone numbers. Drafting only: no skill sends anything.

9 skills · 1 MCP server

SDD WORK-PIPELINE Agent — Requirements analysis & development 6-agent full pipeline with DAG-based orchestration and sliding window context management

4 skills · 6 agents

Reads how much of your 5-hour and weekly Claude Code limit is left, converts it into turns of headroom rather than a percentage, and plans the work to fit inside it. Detects plan tier, splits spend by model and project, and adds a status line.

1 skill · 8 commands · hooks

Switch to a fresh Claude chat without losing a single rule, correction or number. /amhandoff writes a complete handoff file, and a code word plus checkpoints warn you when it is time to switch.

8 skills · hooks

Agents for academic, market, and technical research, synthesis, and reporting

9 agents

Multi-agent development system with three specialized agents (PM, Dev, QA) coordinated through checklist methodology based on The Checklist Manifesto and Boeing's checklist engineering

6 skills · 3 agents

GitHub workflow skills for Claude Code - issue creation, PR review triage, backlog summaries/quick-wins, release status, and GitHub Actions authoring. Seven skills; three pure-LLM, four script-backed via an auto-materialized plugin venv.

8 skills · hooks

Commands for orchestrating complex workflows

9 commands

Advanced frontend design plugin with interactive wizard, trend research, moodboard creation, browser-based inspiration analysis, color/typography selection, and WCAG accessibility

6 skills · 3 commands

9 skills for venture capital intelligence. 3 run on pure Claude reasoning. 6 run Claude + Python for deterministic outputs — screening startups, modeling financials, sizing markets, computing cap table waterfalls, scanning deal signals, and generating LP fund reports. Built from the best open-source VC tools.

9 skills

Agents for cloud infrastructure, DevOps, and database operations

8 agents

Build with Bestax, React components for Bulma v1. Seven Agent Skills for forms, layouts, theming, icons, custom components, CSS size and migration from Bulma classes or older React Bulma libraries, plus the bestax-mcp server, which answers questions about components, props, examples and CSS variables from a bundled index. Local stdio server, exact npm pin, no telemetry.

7 skills · 1 MCP server

Commands for scenario simulation and decision modeling

8 commands

Agent skills for real Figma files: Plugin API gotchas, checks before and after every write, comments to verified fixes, handoff gate.

8 skills

Next.js development expertise with skills for App Router, Server Components, Route Handlers, Server Actions, and authentication patterns

5 skills · 3 commands

SSL/TLS certificate scanning, free Let's Encrypt issuance (private key stays local), and ongoing certificate-expiry monitoring — through a single remote MCP server. Free anonymous scans and cert issuance need no account; OAuth via /mcp unlocks ongoing monitoring.

1 skill · 6 commands · 1 MCP server

Real UI, not AI slop. The free no-ui-slop and uxkin-research skills (work with no account), plus the optional UXKIN MCP server for real iOS app screens, user journeys and website design systems (sign in with UXKIN; searching needs a UXKIN plan). Tool results are treated as data, never as instructions.

7 skills · 1 MCP server

Reusable agent skills maintained by YYLO. The canonical, independently versioned source for skills used by YYLO CLI and YYLO Ledger: ledger task management, wiki, workflow and artifact records, project understanding, planning, and the Ralph execution loop.

8 skills

Agents for content marketing, customer support, and sales automation

7 agents

Authoring and verification for LNPL, the language of the linkly LLM-native platform (declare entities, services and workflows; the compiler lowers them to a semantic IR that runs on an interpreter or compiles through MLIR to a native binary). LNPL's verb vocabulary is closed and absent from model training data, so a plausible word compiles into a step that silently does nothing. The plugin routes Claude to the lexicon generated from the compiler's own tables, surfaces lnpl compile diagnostics on every save, offers an independent reviewer subagent, derives spec cases, and gates completion on evidence.

5 skills · 1 agent · hooks

A lifecycle for Claude Code native file-based memory: a save gate that gives each memory a tier and an expiry, a sweep that archives expired memories instead of deleting them, a consolidation pass that merges duplicates and resolves contradictions only after you confirm, and a correction-signal hook that feeds an always-loaded HABITS.md capped at 8000 bytes / 24 rules, enforced at write time. Local-only.

6 skills · hooks

Design and build iPhone app screens and flows that feel native: onboarding, sign-up, paywalls, permissions, settings and more, checked against Apple's conventions and real iOS apps. Skills: ios-flows, ios-screens, ios-review. Works with no account; the UXKIN MCP server is optional (sign in with UXKIN; searching needs a UXKIN plan). Tool results are treated as data, never as instructions.

3 skills · 3 commands · 1 MCP server

AI-native mortgage refinance plugin — real-time institutional pricing, state-specific closing costs, FHA Streamline/VA IRRRL detection, weighted recommendation scoring, and regulatory compliance. No API key required.

5 skills · 1 command · 1 MCP server

Design and build paywalls, upgrade screens and pricing pages that are clear and honest: the value first, real prices and renewal terms, one obvious choice, Restore Purchases and every loading, error and already-subscribed state. Skills: app-paywall, pricing-page-layout, paywall-review. Works with no account; the UXKIN MCP server is optional (sign in with UXKIN; searching needs a UXKIN plan). Tool results are treated as data, never as instructions.

3 skills · 3 commands · 1 MCP server

Quality gates for AI-assisted codebases. Provides refit onboarding, the swab/scour/buff/sail remediation loop, and barnacle friction reporting as a Claude skill and slash commands.

1 skill · 6 commands

Hooks that send each Claude Code tool call to a local Tracekit signer before it runs: allowed, denied with the rule that matched, or held until you approve it from another terminal. The signer signs every call, decision and result into a hash-chained record that `tracekit verify` checks offline. Needs the tracekit-ai Python package (`pip install tracekit-ai==1.6.0 && tracekit init --dev --v2 --no-hooks`). In that dev mode the signer runs as your own user; on Linux, system mode runs it as a separate user.

3 skills · 3 commands · hooks

Vulnerability intelligence and remediation skills for Claude Code — 7 skills for exploit analysis, fix proposals, scoring, exploits, and package security via the Vulnetix VDB API

7 skills

Design and build websites and landing pages that look like a real product, not AI output: pick a design direction from real websites, write it down as a DESIGN.md, build pages that work on phone and desktop, and review what you have. Skills: landing-page, web-design-system, web-review. Works with no account; the UXKIN MCP server is optional (sign in with UXKIN; searching needs a UXKIN plan). Tool results are treated as data, never as instructions.

3 skills · 3 commands · 1 MCP server

Claude Code efficiency and hygiene toolkit. Surfaces token spend across the 5h / 7d / Sonnet-7d billing windows with per-model / per-agent / per-skill attribution, and audits your effective Claude Code configuration (custom + plugin agents and skills) for overlap and conflicts. Ships an interactive dashboard, a conversational usage-analysis skill, and a structured config-audit skill.

5 skills · hooks

Helps Claude make deterministic, auditable choices about which agent and skills to use for a given task — replacing prose-scanning agent/skill selection with a typed scoring kernel that returns one of seven typed decisions with confidence, rationale, and alternatives.

5 skills · hooks

Commands for optimizing build, bundle size, and performance

6 commands

Commands for initializing and setting up new projects

6 commands

Query the KEGG bioinformatics database — pathways, genes, compounds, drugs, enzymes, diseases, reactions, and more via 34 read-only MCP tools. No API key required.

1 skill · 3 commands · 1 agent · 1 MCP server

Turn Claude Code into a visual creative assistant — AI image generation with creative workflow orchestration, prompt engineering, parallel multi-direction output, and 1,300+ curated inspiration library

2 commands · 3 agents · 1 MCP server

TraderSpy — Read-only smart money tracking, AI trading signals, and elite trader leaderboard for crypto futures. Track whale positions across Binance, Hyperliquid, Bybit, and OKX. No order or withdrawal tools.

6 skills

Agents for multi-agent orchestration, MCP tooling, and agentic workflows

5 agents

Expert agents for cryptocurrency trading, DeFi strategies, and market analysis

5 agents

Five-agent meta-orchestration kit from operator-kit: design, build, scout, research, and critic specialists

5 agents

Watch every Claude Code, Codex, Mistral Vibe and Cowork session on this computer, sub-agents included, as a live animated team call: who is working, stuck on a long command, waiting on another agent, or waiting for you, with tokens per agent. Two skills: /agent-standup:open starts the local app (`npx -y @double_geste/[email protected] --open`, Node 18+, localhost:4747) and /agent-standup:status gives a one-line summary. Reads local session files only and sends nothing by default.

5 skills

Cuts Claude Code's token use on three axes: terse answers that keep every decisive fact, smallest-code-that-works rules, and a PostToolUse hook that trims oversized tool output before it re-enters the context. Benchmarks with raw transcripts, including the runs it loses, are published in the repository.

4 skills · hooks

deja-vu memory for Claude Code: your own past sessions, from every coding agent on the machine, recalled before you ask. Indexes the session files 44 agents (Claude Code, Codex, Cursor, opencode and more) already wrote to disk into a local BM25 index, and adds an MCP server, a /deja command, a skill and hooks that recall at session start, on each prompt and before edits or commands. No LLM, no embeddings, no network. Needs the deja binary (brew, npm, scoop or winget).

2 skills · 1 command · 1 MCP server · hooks

Fix production bugs with LogBrew error tracking, performance monitoring, logs, and product analytics. Connects the hosted LogBrew MCP server over OAuth and adds 4 skills: debug a production error, instrument an app, read product analytics, and check a release. Reads only the projects you approve. Every step that creates, installs, records, or changes something shows the exact command and waits for your yes. Generous free tier, no card needed.

4 skills · 1 MCP server

MoltNode by Chord: two hosted MCP servers and three skills, no API key. MoltNode reads 31 blockchains over JSON-RPC; Chord returns best-route DEX swap quotes and unsigned swap transactions (with the approval and a simulation) on Ethereum, Base, Arbitrum, BNB Chain, Avalanche, Monad, Robinhood Chain and Arc, plus token boards, candles, wallet PnL and smart money. Non-custodial: the user's own wallet signs every transaction, after the user confirms the quote in a later message. Skills: moltnode (reads and swaps end to end), chord (trading and market data), chord-charts (candlestick SVG charts for tokens on 8 EVM chains and Solana).

3 skills · 2 MCP servers

Pharon, the warning light for your stack: keeps a local inventory of your packages, runtimes, container images, CI actions and services, and alerts your team when a compromised release, an exploited vulnerability, a vendor breach or an end-of-life date reaches it. Scans run on your machine and only the inventory is uploaded, after you approve the first upload in /pharon:setup. Needs a Pharon account; a 7-day free trial is available.

4 skills · hooks

Sequential three-tier code review (Haiku surface → Sonnet logic → Opus deep), restart-on-fail. Battle-tested across hundreds of solo issues.

1 skill · 1 command · 3 agents

Ravenpost — draft, preview, schedule and publish social media posts to Instagram, TikTok, X, LinkedIn, Threads, Bluesky, YouTube, Pinterest and Telegram from Claude, and read followers and engagement. Every post is previewed first; a schedule is made only after you say yes to the accounts, caption and time, and a publish waits until you confirm it.

4 skills · 1 MCP server

Train, compile, deploy and profile neural networks on Texas Instruments Edge AI processors (C7x/MMA with TIDL: TDA4VM, AM68A, AM69A, AM67A, AM62A). Five skills cover TIDL compilation and int8 calibration in Docker, host-vs-board output checks, edgeai-gst-apps configs and per-layer profiling. Asks before any board-state change.

5 skills

Agents for business analysis, financial modeling, and KPI tracking

4 agents

Adversarial verification primitives for Claude Code: catches silent skipped restarts, probability-stacking errors in reports, and lost-PID daemon spawns on remote hosts.

3 agents · hooks

Commands for designing and documenting REST and GraphQL APIs

4 commands

Commands for loading context and priming Claude for specific tasks

4 commands

General-purpose utility commands

4 commands

Commands for security auditing and vulnerability scanning

4 commands

Codex-style computer use for Claude Code. Drives Windows and macOS apps through the accessibility tree instead of screenshots, so it targets exactly, survives display scaling, and costs about 20x fewer tokens. It can also work on the same desktop while you keep working, telling your input apart from its own.

1 skill · 1 command · 1 MCP server · hooks

MiCA compliance evidence, stablecoin risk scoring, and tamper-proof audit trail. 27 MCP tools with ES256K-signed responses for regulated tokenized markets.

1 skill · 3 commands

Keeps a MacBook awake (lid closed, even on battery) while Claude Code works, then lets it sleep when it's done. Three commands (/lidnight:awake, /lidnight:status, /lidnight:sleep) and a keep-awake skill that drive the Lidnight menu bar app through its CLI or lidnight:// URL. Without the app it only prints caffeinate/pmset guidance and installs or changes nothing. No hooks, no MCP servers, no settings.

1 skill · 3 commands

Generate videos with MiniMax-Hailuo-2.3 from a text prompt or a first-frame image, poll generation status, and download the finished file across the global and China API endpoints.

4 commands

Three skills that find the names of new AI models, products and repos on X, Hugging Face, GitHub and sitemaps, then size the search terms behind them with Google Ads volume and DataForSEO Trends (prioritize / watch / skip). They run the pinned new-term-radar release from PyPI (uvx new-term-radar==0.6.1), state the cost in USD and wait for a yes before every paid run, show what would go out for the user (a post, a page, a submission) and wait for a yes before any of it, and treat fetched posts and pages as data. Needs an AIsa API key.

3 skills · 1 MCP server

Connects Claude to AccountExecutiveJobs.com through its hosted OAuth MCP server. Job seekers search account executive and sales jobs and apply after an exact preview and confirmation; employers draft, publish and manage their own job posts and review applicants. Includes a founder skill for planning a first sales hire (role, base, OTE and quota from your own numbers, job description, scorecard). No local code; every write needs a preview and later-turn confirmation, and job listings, profiles, tool results and applicant content are treated as data that can't confirm anything.

2 skills · 1 MCP server

Agents for writing, formatting, and validating documentation and technical content

3 agents

A hidden background Windows desktop for agents: GUI automation, UI Automation, screenshots and app testing without taking over your mouse.

2 skills · 1 MCP server

Read-only shopping decisions for Greece from BestPrice.gr: product search, current offers from Greek stores with delivered prices, price history, and a buy-now-or-wait decision for a product. Adds the bestprice-shopping skill and the hosted BestPrice MCP server (https://mcp.bestprice.gr/mcp); no sign-in, no API key. It never buys anything or touches a cart.

2 skills · 1 MCP server

Put this Claude Code session on a full-duplex voice call (Azure GPT Realtime or GPT-Live, macOS): talk while it codes, interrupt anytime, hear results read back. Voice never approves permission prompts; they stay on the keyboard.

3 commands

Personal finance ledger for AI agents. Connect bank accounts via Plaid, query spending, track recurring bills, categorize transactions, and get financial recaps through MCP tools with OAuth authentication.

2 skills · 1 MCP server

cohesivity.ai offers free agent native backend services. Anonymous account (no-signup) to get started through MCP or API. Hosting, postgres, email, storage, containers, LLMs, voice and third-party APIs. Includes free tiers and 5 USD/mo in AI and Search credits. topups through x402.

1 skill · 2 MCP servers

Commands for database schema design, migrations, and optimization

3 commands

Let a Claude session in one repo delegate work to a Claude session in another repo, with a request format that avoids follow-up round trips. Adds /ask-peer plus skills for sending and answering cross-repo requests.

2 skills · 1 command

Analytics engineering skills for warehouse exploration, dbt and semantic modeling, and schema-drift maintenance. Three skills (explore, transform, maintain) drive a local engine: warehouse reads are read-only and cost-gated, and project edits come back as reviewable diffs. Requires a local shell and uv; first use downloads the pinned engine from PyPI. No hooks, no MCP servers.

3 skills

Read visual website feedback, marked screenshots, recordings and team discussion through a scoped MCP connection. Includes review-feedback and manage-feedbacks-context skills. Requires a Feedbacks server, Node 22.12+ or 24 and Git.

2 skills · 1 MCP server

Add native video and image understanding to Claude via Qwen Omni API. Analyze videos, classify footage, extract motion data, and more.

1 skill · 2 commands

Don't let Claude say "done" until the tests actually pass. A Stop hook runs the repository's real test, typecheck and lint commands and blocks the turn until they pass; a PostToolUse hook warns when an edit weakened a test; a skill teaches Claude to verify before claiming completion. No LLM calls, no network.

2 skills · hooks

Three Claude Code skills for designers who build in Spanish: UI copy review in neutral Spanish with tú (voseo, vosotros, leftover English, title case, missing ¿¡), WCAG 2.2 contrast measured on the project's color tokens in light and dark mode with the nearest passing color, and a basic SEO pass over a built or live site. Three bundled Node 18+ scripts that read files and write only a report when passed --json; seo-basico fetches only the URL the user gives. No hooks, no MCP servers, no settings.

3 skills

Skills for working with Obsidian files including Markdown with wikilinks/embeds/callouts, Bases for database views, and Canvas for visual diagrams

3 skills

Multi-agent orchestration and project-memory system for Claude Code. One command turns a repository into a wired multi-agent environment: a project-specific agent roster, a documentation and session-summary tree, a tested script library, and enforcement hooks that hold size and stamp discipline at the harness layer. Re-runnable, so the setup grows with the codebase.

2 skills · hooks

Pixel art and sprite animation from editable JSON recipes, with particle and fluid effects; exports PNG/GIF/APNG and atlases.

2 skills · 1 MCP server

Find what's going viral in your niche, write your version in your own voice, and schedule it to X, LinkedIn, Bluesky, Mastodon, YouTube and TikTok through your Postel account. A remote MCP server (OAuth) plus a writing skill.

2 skills · 1 MCP server

Your versioned prompt library inside Claude Code: fetch, fill and save prompts from your PromptLikeEngineer account, run them from the / menu, and get a local hint when a saved prompt fits what you ask. It can save prompts, add new versions and move prompts between repos and folders; there is no delete tool. Needs a PromptLikeEngineer account; a free plan is available.

1 skill · 1 MCP server · hooks

Shows Claude Code sessions as villagers in the Wudland iPhone app: working, waiting, done or failed. Updates are encrypted end to end over an ntfy relay; no account. The iPhone app is coming to the App Store.

2 commands · hooks

Specialized agents for blockchain development, smart contracts, and Web3 applications

2 agents

Agents for UI/UX design, accessibility, and user experience optimization

2 agents

The session-lifecycle commands from aigent-OS, the open-source Claude Code operating system: /open (boot the session — load memory from the vault, surface what changed and what's blocked) and /close (bank the session back to memory so the next /open resumes with full context). Both skills declare scoped allowed-tools and run a vault-root preflight that verifies the configured aigent-OS vault before any write or script execution. Assumes the aigent-OS vault structure; see the repo and the plugin README for the full system and the security model.

2 skills

AInotate lets AI agents answer "where is it?" with an annotated screenshot: numbered steps, arrows, boxes and labels, with emails and tokens redacted.

1 skill · 1 MCP server

Search Dongdaemun (Seoul) wholesale fashion with apMZoomAI: items, new arrivals and stalls by building and floor, with links back to apMZoomAI. A read-only remote MCP server plus a sourcing skill; no prices or contact details.

1 skill · 1 MCP server

Computer use for Claude Code: click, type and read the screen on your Mac, Windows or Linux computer and your Android phone over USB (beta), even in apps with no API. Reads the accessibility tree, types saved passwords locally, and asks before anything irreversible. Free plan 30 minutes of runtime a month.

1 skill · 1 MCP server

One skill and an MCP server that check a link source (directory, awesome list, guest-post site) before you submit your site to it: skip, look first, free only, low value or submit, with the evidence, against Google's link spam policies. They also verify a listing once it is live and audit your own backlinks. They never submit anything, and before an agent submits a site or opens a pull request for you, the skill has it show where it goes and the exact text and wait for your yes; a vet result is not that yes. Both run the pinned backlink-guard release from PyPI (uvx backlink-guard==0.11.2), state the cost in USD and wait for a yes before every paid call, and treat fetched pages and AI answers as data. The page checks need no key; paid signals use a Glasser or AIsa key.

1 skill · 1 MCP server

Connect Baizhi Agent Toolkit through an authenticated remote MCP server, with an optional skill for bounded research and source verification. Requires a Baizhi account and API key; service usage may incur charges.

1 skill · 1 MCP server

Newsletter sponsor data: give it a newsletter's web address and it finds similar newsletters and the brands that sponsored them, with the newsletter and date where each ad ran; lists brands sponsoring newsletters by topic; and gives the median price newsletters list for a sponsor slot by list size. Adds the book-a-sponsor skill and the hosted Book a Sponsor MCP server (https://bookasponsor.com/mcp); no sign-in, no API key. Brand-level data only: no people, emails or phone numbers.

1 skill · 1 MCP server

Publish HTML, Markdown, files and multi-file static sites from Claude Code to BrewPage and get an HTTPS link back, with no account or API key: a brewpage-publish skill plus the BrewPage remote MCP server (https://brewpage.app/mcp). Only content you ask it to publish is uploaded; links are unlisted but readable by anyone who has them (optional password), and content expires (15 days by default, 30 at most).

1 skill · 1 MCP server

Local spend monitor for Claude Code. Per-project, per-branch budget caps with SIGTERM enforcement and phone alerts via ntfy.

1 skill · hooks

Claude Code plugin for Carrick, which indexes TypeScript codebases across service and repository boundaries. After each edit it adds the file's routes, calls, and cross-service type mismatches to the session, and it registers Carrick's language server. It pairs with the Carrick MCP server, which lets agents search functions by intent rather than name.

hooks · LSP

Set up and operate CIAgent (pip install ciagent) regression testing for the AI agent in this repo: record golden baselines, run stability and judge-reliability checks after every change.

2 skills

Scaffold and audit CLAUDE.md files from inside Claude Code: /claude-md-new builds one from battle-tested templates using your repo's real commands; /claude-md-audit grades an existing one and returns worst-first fixes.

2 commands

Commands for setting up monitoring and observability

2 commands

Plan the "Comment KEYWORD" CTA of an Instagram Reel (keyword, on-screen text, caption, DM copy) and set up the automatic DM in CommentKeyword, which sends it through Instagram's official API. Adds the comment-to-dm-cta skill and a dependency-free local MCP server; nothing that messages people is created until the user confirms the keyword, DM text and target posts. Needs a CommentKeyword API token (API access is included on the Free plan).

1 skill · 1 MCP server

Which subreddits AI answers cite: describe a product or niche and it returns the subreddits where it fits, with how often AI answers (OpenAI models, Perplexity, Gemini) cited each one and its member count; ask about one subreddit and it returns its citations by engine, the kinds of questions those answers were about, and its most-cited threads. Adds the communitymentions skill and the hosted CommunityMentions MCP server (https://communitymentions.com/api/mcp); read-only, no sign-in, no API key.

1 skill · 1 MCP server

Read-only conference and trade show data: which companies sponsor and exhibit at an event, the conferences a company sponsors or exhibits at, event dates and places, and the sponsorship packages organizers publish. Adds the conferencedb skill and the hosted ConferenceDB MCP server (https://conferencedatabase.com/api/mcp); no sign-in, no API key. Company-level data only: no people, emails or phone numbers.

1 skill · 1 MCP server

Publish a page Claude wrote, or an HTML file in your repo, at a convika.com address or on your own domain. Convika keeps signups from the page's form and counts visits, traffic sources, and button clicks. Sign-in is OAuth in the browser, so Claude never holds an API key. The publish-page skill shows a private preview and asks before it publishes.

1 skill · 1 MCP server

Stop/PostToolUse hook: after a code edit, Claude Code cannot end its turn until a recognised test command has run and passed after the last change. Piped, masked or failing runs do not count; after two reminders it lets the turn end with a warning. A repo's .proof-of-done.json can only add checks; waivers live in the user's ~/.claude/proof-of-done.json. Standard-library Python, no network.

1 skill · hooks

The inverse of /grill-me: an adaptive tutor that teaches you anything — spaced repetition, retrieval practice, and a memory that persists between sessions.

2 skills

US stock market research through the hosted Equibles MCP server: SEC filings with full-text search, XBRL financial statements, earnings call transcripts, insider and congressional trades, 13F institutional holdings, short interest, end-of-day prices, option chains and macroeconomic data. Sign in over OAuth from /mcp; free tier of 100 requests a day, no credit card.

1 skill · 1 MCP server

Strategic foresight intelligence using IFTF methodology. Two prediction modes: Soft Predict Future (Claude-native skill, instant, claude.ai + Claude Code) and Hard Predict Future (deterministic 12-step pipeline, Python-computed — Claude Code only). Both output four independent futures (Probable/Plausible/Possible/Preferable) with per-stakeholder conditional analysis. Try: Will India become the global AI leader? Will OpenAI or Anthropic win?

2 skills

Free Google Search Console MCP with Bing Webmaster Tools. Connects the hosted CrakSEO Search Console server over OAuth and adds a skill: your own sites' search queries, clicks, impressions, CTR and average position, URL inspection, sitemaps, and Bing traffic and crawl issues. Every tool is free. Submitting a sitemap is the only write, and the skill asks before it.

1 skill · 1 MCP server

PreToolUse Bash guard for Claude Code that still holds under --dangerously-skip-permissions. Blocks pipe-to-shell installs, disk wipes and fork bombs; asks before rm -rf, force-push, git reset --hard, DROP/TRUNCATE, kubectl delete, cloud deletes and secret-file reads. 15 rules, each off/ask/block; a project's own config can only tighten them, loosening comes only from your config outside the project. Plus a PostToolUse audit log with secrets redacted and /guardrails:self-test to see every decision without running anything.

1 skill · hooks

One skill for running Claude Code on Herdr, the terminal agent multiplexer. spawn-session starts a named, detached Claude Code Remote Control session for any project directory in its own Herdr pane, so it appears in the Claude mobile app, and confirms on the pane screen that each prompt landed.

2 skills

HookRadar creative intelligence workflows for Codex, Claude, and MCP-capable agents: competitor ads, organic videos, AI creative analysis, reports, and free first-pass competitor research.

1 skill · 1 MCP server

Keep a MacBook with Apple silicon awake with the lid closed while long agent tasks run. Adds the keepmacawake skill and a read-only session-start hook: the agent starts a timed session of the Keep Mac Awake app before long work, stops it after, and tells you when the app needs something from you (AC power or a purchase). Needs the Keep Mac Awake app on macOS 26 or later, installed separately: a free 3-day trial, then 9 USD once per Mac. The plugin never uses sudo, downloads nothing and makes no network connection.

1 skill · hooks

Search and read Korean law through the official Ministry of Government Legislation (법제처) Open API: statutes, precedents, administrative rules, local ordinances, treaties and legal interpretations. Also checks legal citations in Claude's answers against the source for existence and content. Runs the korean-law-mcp MCP server via npx and needs a free 법제처 API key (LAW_OC), entered at install.

1 skill · 1 MCP server

Search, price and run 2,500+ data API endpoints (work emails, company and people search, Google results, web pages) with one prepaid balance. The agent sees the price before a call runs, and a failed call costs nothing. Remote MCP server with browser sign-in plus one skill.

1 skill · 1 MCP server

Trading card market prices and grading ROI through one remote MCP server. Raw and graded prices for 1.5M+ Pokémon, Magic, Yu-Gi-Oh!, One Piece, Lorcana and sports cards, PSA 9 and PSA 10 premiums, net profit after grading fees, price history, 30-day movers and priced set checklists. Read-only, no account, no API key.

1 skill · 1 MCP server

Open Claude Mission Control, a local read-only dashboard for every Claude Code session: who is waiting on you, cost per session and commit, and a usage-limit forecast. Adds a statusline.

2 skills

Inspect and interact with Android and iOS apps through the mobile-debug-mcp server and a usage skill.

1 skill · 1 MCP server

Two skills for generative UI in Angular with ngx-json-render, the Angular renderer for Vercel's json-render: define a component catalog and registry, render model-written JSON specs with <json-render>, stream them with signals, check specs and test components, and use the ready Angular Material catalog.

2 skills

Connect Claude to Pine AI to handle real-world tasks and follow up on progress and results.

1 skill · 1 MCP server

Submit, revise and track research papers at Publish.fun, an AI-reviewed research journal, and cite published papers. Bundles the journal's remote MCP server and its agent skill. Reading and citing need no key; submitting needs an API key from an ORCID-verified account.

1 skill · 1 MCP server

Lets Claude Code start other AI coding CLIs (Claude Code, Codex, Kimi, Qwen, OpenCode) in their own tmux windows, send them prompts, read their output and stop them, through the reevesagents MCP server. Ships one skill and the MCP config; needs the reevesagents npm package and tmux.

1 skill · 1 MCP server

Two skills for rendering agent output safely: replace innerHTML with render-policy's sanitize-into-a-fragment renderers for plain DOM, React and Angular, with image, link and host policies, streaming and ESLint rules; and lint the CSP and tool metadata an MCP App declares before you publish it.

2 skills

SEO data for Claude Code: keyword research with search volume, difficulty and CPC, live Google and Bing SERPs with AI Overview sources, rank tracking and competitor keyword gaps, backlinks and referring domains, site audits with Core Web Vitals. Connects the hosted CrakSEO SEO MCP server over OAuth and adds a skill. Paid tools use prepaid credits; the skill states the cost and asks before every paid call, and never buys credits.

1 skill · 1 MCP server

Describe your app in plain English — Shipwright builds, tests, and deploys it autonomously. 9-phase pipeline, 4 stacks, enterprise-grade safety hooks.

1 command · 1 agent

Stop hook for Claude Code and Codex. When the agent ends with "all tests pass", "the build succeeds" or "lint is clean", it checks the session transcript for a matching command that ran after the last edit and exited 0, and blocks the stop if there isn't one. One file of standard-library Python, no LLM, no network.

1 skill · hooks

Browser automation in Claude Code through Skyvern Cloud: open pages, fill forms, extract structured data, take screenshots, and run repeatable workflows. Sign in with OAuth from /mcp. Runs can use Skyvern credits.

1 skill · 1 MCP server

Run a GPU job under a hard dollar cap and collect the files it writes.

1 skill · 1 MCP server

Run AI QA tests on your web app from Claude Code and fix the bugs they find. The testoptim-qa skill re-tests the pages you changed with TestOptim, reads the critical and high issues (steps to reproduce, page URL, screenshot), fixes them and re-tests. It asks before every test run, because runs use your plan's executions. Needs a TestOptim account and an API key, which the plugin asks for at install.

1 skill · 1 MCP server

Pre-action gates for AI coding agents. Type 👎 on any agent action — ThumbGate captures it, distills a lesson, and blocks the pattern from repeating. 33 pre-action gates, budget enforcement, self-protection, and NIST/SOC2 compliance tags.

1 skill · 1 MCP server

Find agent skills for a task, read side-by-side proofs of what they change, and safety-check any skill (even straight from a GitHub link) before installing. Adds the find-proven-skill skill and the hosted Use This Skill MCP server; no sign-in, no API key.

1 skill · 1 MCP server

Operate a real iPhone from Claude Code through a local MCP server in the Zeptap Mac app. Claude sees the screen over USB and taps and types over Bluetooth, with nothing installed on the phone. Tools include screenshot, find_text, tap, tap_text, type_text, open_app and unlock. The Mac app is closed source and paid (macOS 14+), with the first hour free.

1 skill · 1 MCP server

Creates hand-drawn animated technical diagrams as self-contained HTML and seamless-loop GIFs, with motion that follows branches, merges, and feedback loops in the system topology.

1 skill

Runs other coding-agent CLIs (Codex, opencode, Pi, Antigravity, another Claude Code) as background subagents through one agent-send command. Each delegated conversation has a name, so a follow-up resumes the same backend session with its context. Roles map to backends in user config, and each agent logs its spec and notes to a per-session file. One bash script and a skill; no hooks, no MCP servers.

1 skill

Ask real executives, operators and domain experts a question through Instant Expert and get a written or voice answer or a short call. It finds people by description or reaches a named person, and the user approves every live order. Starts in a free test mode with no account; you pay only if someone answers.

1 skill

End-to-end encrypted, quantum-resistant chat between Claude Code users: direct chats, groups and read receipts in a pane, with no server to run.

1 skill

Keeps Claude Code sessions on AWS Bedrock alive through credential expiry. Scripts for awsCredentialExport and awsAuthRefresh, added to settings.json by hand, read fresh credentials from ~/.aws/credentials and can re-run your login. A UserPromptSubmit hook warns when expiry is near or past. Works with saml2aws, gimme-aws-creds, aws-google-auth and any tool that writes to ~/.aws/credentials.

hooks

Real-time statusline HUD for Claude Code - displays context usage, tool activity, agent tracking, and todo progress

1 command

Generate and edit AI images and videos from inside Claude Code through one remote MCP server. Nano Banana 2, GPT Image 2.5, Seedream 4.5, Flux 2 Pro and Z-Image for stills; Veo, Kling, Seedance, Grok Imagine and MiniMax for clips. quote_generation prices any call before you spend, and failed generations are refunded automatically. OAuth via /mcp; no API key to paste.

1 MCP server

Native desktop notifications when Claude Code needs your attention

hooks

Rewrites your request into an optimized, reasoning-maximizing prompt (structured context, specificity, meta-instructions for extended thinking, skip-comments that preserve quoted text), then immediately carries it out. Ships the /p command.

1 command

Lets Claude check its own claude.ai plan quota: percent used of the session and weekly limits (including per-model weekly limits), when each resets, and where usage is on pace to be at reset. One skill and a short Python script that reads Claude Code's stored login token and calls the same usage endpoint as /usage. Claude Code only; no hooks, MCP servers, or telemetry.

1 skill

While Claude works, Clawd (Claude Code's mascot) acts out the spinner's -ing word and the running tool on a little stage above the prompt: 26 animated scenes, wandering the band, changing scene every 30 seconds on long turns.

1 skill

Cheeky Cockney coding mode: terse replies, the smallest correct change, plain English for anything risky.

1 skill

Commands for automating repetitive tasks and workflows

1 command

Commands for game development workflows

1 command

Commands for migrating JavaScript projects to TypeScript

1 command

Deep-reading assistant: five modes (quick, deep, knowledge-map, Feynman, full book) that extract core claims, argument structure, and key evidence from books and articles into claim-evidence-data reports with four confidence levels; batch comparison, budget preflight, and background-job progress transparency; exports Markdown, FreeMind/XMind mindmaps, and HTML reports.

1 skill

Design and generate .envelope.json AI agent team definitions — the open standard for multi-agent teams with hierarchy, access policies, human-in-the-loop gates, and cron schedules. Validates against schema.openenvelope.org.

1 skill

Interview command for fleshing out big feature plans and specifications

1 command

Kernel Field Guide: Linux kernel development the way experienced kernel engineers work: driver code and teardown order, devicetree bindings and DTS, Kconfig, debugging by symptom, KUnit and kselftest, patch review and submission, regressions and bot reports, and stable, LTS, and vendor kernel trees.

1 skill

Read, create, fill, patch and compare Korean documents in Claude. One skill that runs the kordoc CLI via npx to convert HWP 3.x/5.x, HWPX, HWPML, PDF, DOCX and XLS/XLSX files to Markdown, fill form templates while keeping their formatting, and generate HWPX documents from Markdown with Korean government document presets. No Hancom Office needed.

1 skill

Finds B2B companies that fit your ideal customer and show a buying signal now, with a source link and an exact quote for every claim. Anything it can't check stays "unknown" instead of being guessed. A deterministic, standard-library Python script does the scoring, and hiring is checked on 7 public job-board APIs (Greenhouse, Lever, Ashby, SmartRecruiters, Recruitee, Breezy HR, Personio). No API keys.

1 skill

The LNPL compiler (linkly LLM-native platform) as MCP tools instead of a shell command. lnpl compile writes diagnostics to stderr and exits 0, so a caller that checks the exit code misses every warning, and an unknown-verb warning means a step that does nothing at runtime. lnpl_compile, lnpl_spec and lnpl_kb_route return diagnostics, spec results and knowledge-base routes as structured records, calling the compiler in-process with no PATH lookup. The compiler is loaded only from a path you configure (LNPL_IMPL_PATH) or an installed package, never from the project directory.

1 MCP server

127 auto-loading skills for Claude Code with localhost dashboard, 3-agent orchestration, session diary, and token tracking. 85 free, 42 Pro.

1 skill

Generate music with MiniMax models across the global and China API endpoints.

1 command

Prepare flight comparisons with airport lookup, validated trip inputs and ready-to-open search links. Free hosted MCP; no API key, live fares or bookings. MIT connection plugin; private hosted backend.

1 MCP server

Share-image sets for Xiaohongshu and X that look native: App Store poster layout, real screenshots, faithful app-UI mocks, charts instead of tables, covers and a series layer for building in public, rendered from HTML templates to PNG.

1 skill

Production-ready standards for a scalable, unified FastAPI + RAG + Agentic AI backend on AWS ECS Fargate: project structure, async execution, Pydantic and structured outputs, API conventions, guardrails, AI evaluation (DeepEval, Ragas), API security and LLM/agent security.

1 skill

Security hook that blocks destructive commands and file operations outside the project directory

hooks

Take an app or product from idea to launch, taste-first: positioning, naming with domain and trademark checks, reference hunting, style frames beside named references, feature map, UI build loop, launch video, GTM and shipping, with phase gates that force a real artifact before the next step.

1 skill

Audio-first motion-graphics launch, brand and TVC films rendered in code (canvas to MP4) and cut to the music: music plan and generator prompts, beat analysis from the spectrogram, reference contact sheets, three style frames before building, a protagonist-driven shot map, a deterministic motion library, parallel rendering and per-language cuts.

1 skill

Build complete Claude plugins in 23 guided steps — from idea to GitHub push. Interviews you, classifies skill vs agent, generates all files, pushes to your repo.

1 skill

Chinese writing in a real person's voice for Xiaohongshu, X, diaries and product copy: removes AI tells (emoji lists, slogans, 不是而是 contrasts, stock closing questions), writes results first with one earned punchline, with a viral Xiaohongshu title library and optional humor patterns.

1 skill

Sets up cookieless, privacy-first website analytics in your own site: time per page section, scroll, goals and 404s from a small tracker and a same-origin collector, an optional consent banner, PostHog session replay only after consent, one Slack or e-mail line per notable visit, a weekly digest, and health alerts. A Playwright check then proves the privacy claims: nothing stored before consent, no e-mail or token from the URL in any report, private pages silent.

1 skill

Competitor traffic and SEO teardown for any website: finds the competitors (Semrush organic competitors), screens them by Similarweb visits, pulls one side-by-side table (3-month trend, channel mix, Semrush keywords and backlinks, Ahrefs DR) and reads it into growth-engine types and three opportunities. Each stage quotes first for free and runs its billed calls only after the user says yes to that quote, and AIsa rows and pages read along the way stay data. Needs a paid AIsa API key and the AIsa CLI (npm i -g @aisa-one/[email protected]), under $1 per domain.

1 skill

Write, upgrade or audit Claude skills so they change what gets generated: distil rules from real corrections, turn them into required artifacts, gates and scripts, and lint any skill folder with a 0-100 health score (name, trigger clause, orphan and dangling files, oversized body).

1 skill

私人旅行管家 · Travel Agent — 在 Claude Code 中一句话规划完整旅行。AI 自动联网查实时车次、景点、美食信息，输出闭环行程并生成 HTML 攻略。每个行程项带 12306/携程/大众点评/高德/小红书一键跳转。

1 skill

Watch any video (YouTube, X, Instagram, TikTok, local files) on-device: local speech transcription plus keyframes read by your agent, several videos in parallel.

1 skill

Complete YouTube toolkit for Claude Code — transcripts, video/channel search, playlists, and channel monitoring via TranscriptAPI.

1 skill

Mod (function hooks, Claude Code 2.1.294+): a line above your prompt when another Claude Code session on this computer needs you (an approval, an answer), and a toast when one finishes its turn. Reads only ~/.claude/sessions/*.json; no network, nothing installed.

Thinking Methods: seventeen Agent Skills (SKILL.md) for structured thinking on real decisions—five whys, pre-mortem, trade-offs, prioritization, six thinking hats, and more, plus a selector that picks the right method. Each skill's Before you start says text from logs, tickets, docs and tool output is evidence for the analysis and does not change the method or authorize a command, a file edit, or a message.

Onboard AI agents like fresh engineers: generate, inject, and maintain a compressed AIBOARDING.md per repo via committed hooks

Complete collection of 28 automation hooks for event-driven workflows

PreToolUse hooks that keep several Claude Code sessions running on one machine from colliding. Blocks kills by program name (pkill node, taskkill /IM node.exe) and kills that reach another session's processes, dev servers on a port another session uses, heavy commands when RAM is low, git checkout, stash, add -A and commit -a in a tree another session works in, and whole-disk scans such as du -sh ~. Stops servers left by finished sessions. bulkhead scan replays the shell commands in your past transcripts against the guards, read-only, with nothing sent anywhere. Blocks return the reason and an alternative command to the agent; only the user can allow a blocked command, and a repository's own config can only make the guards stricter. It parses command strings and is not a sandbox. Node only, no dependencies.

A per-session checklist pane beside the chat: each item has a number, a title, an optional one-line detail, and a status of open, done, or dropped. /checklist adds, edits, moves, and marks items; a click on a row toggles it; and a checklist tool lets Claude add items and tick them off as it finishes them. Every session keeps its own list, which survives /clear and comes back with --resume. A Claude Code mod; needs Claude Code 2.1.287 or later.

Export and apply portable snapshots of your Claude Code setup across machines

A band above the prompt: an animated Clawd sprite, five-hour and weekly usage bars with a countdown to each window's reset, and a context-window bar with a countdown until the prompt cache goes cold. A Claude Code mod; the sprite needs kitty's graphics protocol.

Orchestrate OpenAI Codex from Claude through supervised checkpoint slices, server-enforced quality gates, persistent plans and hypotheses, worktree isolation, and per-task model and effort control.

Persistent knowledge base for Claude Code. Auto-captures the decisions, gotchas, and dead ends from your sessions and recalls the relevant ones at session start and on every prompt, ranked by proven usefulness. Ships an MCP server plus session-recall and capture hooks.

ENCODE Project genomics toolkit — 20 MCP tools + 48 skills for searching experiments, downloading files with MD5 verification, running pipelines, and cross-referencing 14 databases including PubMed, GTEx, ClinVar, GWAS Catalog, and gnomAD

Complete AI coding agent harness for React Native and Expo development — 8 agents, 16 commands, 8 skills, 10 MCP integrations

Human-in-the-loop request queue for autonomous agents: file a blocker (account creation, CAPTCHA-gated step, purchase approval) to a human operator via MCP, poll for the result, keep working. Talks to your own self-hosted Cloudflare Worker.

Nine MCP tools for security audits, web scraping, OG rendering, and funding spreads. Default setup is inspect-only and requires no wallet key. Optional x402 auto-payment requires X402_BUYER_PRIVATE_KEY; use only a dedicated low-balance wallet funded with what you are willing to spend.

Records every Claude Code session live, draws the journey as a map, and coaches the next session.

Read published GoVisit travel information with all available toolkit skills.

Find and inspect destination profiles and their published inventory. Destination updates are planned.

Find and inspect business profiles and their local destination context. Place updates are planned.

Discover destinations and research places, categories, collections, and activities using GoVisit.

Pattern-based PreToolUse-style checks that deny .env reads (Read/Grep/Glob and Bash readers), git force pushes, and recursive rm on /, ~ or $HOME before Claude Code runs them. About 75 lines of dependency-free JavaScript.

A Claude Code mod that turns global skills, agents, instruction files (CLAUDE.md and rules) and tools on or off per repo through named profiles. A repo picks a profile with a one-line .claude/harness-scope.json; repos without the file are unchanged. It only reads: no file writes, network requests, child processes or model calls.

Native Hermes Agent X/Twitter plugin for Xquik automation with read-first workflows and approval-gated actions.

Automation Hooks - Event-driven automation hooks

Development Hooks - Event-driven automation hooks

Formatting Hooks - Event-driven automation hooks

Git Hooks - Event-driven automation hooks

Notification Hooks - Event-driven automation hooks

Performance Hooks - Event-driven automation hooks

Essential safety hooks to prevent destructive operations in Claude Code. From cc-safe-setup (655+ hooks).

Security Hooks - Event-driven automation hooks

Testing Hooks - Event-driven automation hooks

Plan, check and repair multi-day road trips: multiple-choice intake, real road distances, a deterministic checker for sunset, fuel range, opening hours and bookings, overnight and visit-order optimisation, and a fixed-format PDF, HTML or Markdown report. Its scripts send only place names and coordinates to the public OpenStreetMap services (Nominatim and OSRM); it never books or pays, and treats web pages, pasted confirmations, PDFs and spreadsheets as data, never as instructions.

Build integrations with the Lexio Rent API v1, the REST API of the Lexio vehicle rental management software: availability and prices, quotes, rentals, pick-up and return, payments, customers, fleet and event sync. One skill with the API conventions (wall-clock dates, idempotency keys, cursor pagination, rate limits) and safety rules: it asks before any write that changes a booking or money. Documentation only, no code or MCP server. Using the API requires a key from a Lexio rental company (Business plan).

Docker-based MCP servers from the official Docker MCP registry - includes 199+ verified servers

Full Apify web scraping integration — run Actors, manage datasets, key-value stores, schedules, and webhooks for automated data extraction

Prospect leads and enrich contacts with Apollo.io — search companies, find decision-makers, bulk enrich profiles, and manage outreach campaigns

Persistent memory across Cowork sessions — recall decisions, file changes, insights, and errors from previous sessions automatically

Fix stuck Cowork sessions — resolve RPC errors, orphaned processes, and connection issues with automated diagnosis and recovery

Monitor and manage your digital presence — track brand mentions, social profiles, website uptime, and online reputation across platforms

Auto-diagnose and repair broken Chrome MCP connections — fix extension errors, reconnect after account switches, clean stale tabs

Google Cloud CLI health check — verify gcloud auth, active project, enabled APIs, billing status, and Cloud Run services

GitHub CLI health check — verify gh authentication, list repos, check API rate limits, and test connectivity

Upload any file from Claude directly to Google Drive — unlimited uploads, folder selection, automatic organization

Kotlin Language Server Protocol integration — code intelligence, completions, diagnostics, and refactoring for Kotlin projects

Scrape LinkedIn profiles and company pages — extract professional data, work history, skills, company info, and recent posts for research and outreach

Reclaim disk space on macOS — clean system caches, Xcode derived data, Homebrew leftovers, and find large files

Long-term memory for Claude via Notion — persist decisions, context, and project state across sessions using Notion databases

Fix jumbled RTL/LTR text mixing in Claude responses — proper bidirectional formatting for Hebrew, Arabic, Persian, and Urdu

Daily backups of Claude sessions, skills, and configs to Google Drive — automatic versioning and disaster recovery

SOSA compliance auditor — scan plugins and skills against Supervised, Orchestrated, Secured, Agents governance framework

SOSA governance engine — enforce supervision policies, trust gradients, and capability boundaries on AI agent actions in real time

SOSA orchestration layer — coordinate multi-skill workflows with Plan-Act-Verify patterns and structured handoffs between agents

Swift Language Server Protocol integration — code intelligence, completions, diagnostics, and refactoring for iOS and macOS projects

Track time with Toggl — start and stop timers, log hours to projects, generate weekly reports, and manage time entries from Claude

Find and fix token waste in Claude plugins and skills — audit descriptions, skill bodies, MCP overhead, and context bloat. Typical savings: 20-50%

Fix disk-full errors in Cowork VMs and Claude Code sandboxes — clean caches, temp files, and reclaim space automatically

Connect Claude to WhatsApp — search contacts, read conversations, send messages and media files through the WhatsApp Business API

Manage WordPress sites from Claude — create and edit posts, manage users, handle WooCommerce products, and configure settings via REST API

Scrape X/Twitter for insights and generate content calendars — analyze trends, extract engagement data, and plan social content strategy

Transcribe any YouTube video or playlist — no API key required, supports any language, outputs clean formatted text

Zoho Mail health check — verify mail accounts are operational, test send/receive, check DNS records, and monitor deliverability

A band above the prompt that prices the next message before it is sent: a countdown to the prompt cache going cold and what re-warming it will cost, plus session cost, context fill and the 5-hour and weekly limits. A Claude Code mod; it never calls a model or sends data.

A tamagotchi-style pixel pet in a band above the prompt, built on function hooks. It reacts to tool calls, passing checks and commits, and adds a /nibbl command. Uses zero model tokens; syncs event types and times plus a hashed machine id to getnibbl.pages.dev.

Keep the replies, notes and checklists you still need in a side pane next to the chat, so they don't scroll away while you keep working: pin Claude's last reply with /pin, hover a reply and click pin, or ask Claude to pin something. Task lines become checkboxes you tick in the pane, and pins fold into cards. Jump from a pin back to its reply, or fold the pane to a one-line band above the prompt. Session pins come back when you resume; project pins show in every session in the folder. A Claude Code mod; a pin made without a title gets a short one from haiku through your own session.

Shows how long your prompt cache has left and refreshes it before it expires, so your next message after a break doesn't pay to rebuild it.

AI image, video, music, audio generation and LLM chat — 130+ models from 18 providers via RunAPI MCP server.

Claude Code mod that lets you choose the model and reasoning effort of every subagent: ask before it starts, take a recommendation by task kind, or let a saved default replace what the agent requested.

Design, generate, and run multi-agent AI pipelines — scope-aware, multi-pipeline, spec-driven, write/review-isolated, with step-level create/update/delete and integrated audit

All-in-one mod pack for Claude Code: a usage limits and context HUD, a guard with undo for rm -rf and git reset --hard, .env and secret protection, and a receipt for every turn. Ten mods in one plugin; needs Claude Code 2.1.287 or later.

32 expert-level skills for building, shipping, and scaling production software. 29 audit tools (security, code quality, bundle size, SEO/GEO/AEO) close the loop before deploy.

Power your chat with live B2B data to create lead lists, research companies, enrich contacts, personalize outreach, and inspect business signals, tech stacks, events, and website changes. 150M+ businesses and 800M+ professionals via the Explorium MCP.