https://escape9to5.ai /armory A curated library of AI tools, frameworks, and skills — rated, reviewed, and organized by category. en-us Thu, 23 Apr 2026 00:00:00 GMT https://escape9to5.ai /armory/audit-qa/build-verification-gates https://escape9to5.ai /armory/audit-qa/build-verification-gates Thu, 23 Apr 2026 00:00:00 GMT Audit & QA Three gates that catch broken code before your agent commits it. No gate, no commit. When Claude writes code on its own, the real risk isn't bad code. It's bad code that gets committed without anyone noticing. Three cheap gates fix that: 1) typecheck passes, 2) build passes, 3) screenshot proves the page rendered. https://escape9to5.ai /armory/audit-qa/security-sweep-prompt https://escape9to5.ai /armory/audit-qa/security-sweep-prompt Thu, 23 Apr 2026 00:00:00 GMT Audit & QA One prompt that catches exposed API keys, SQL injection, XSS, and auth holes before you ship. Most vibe-coded apps ship with at least one security hole — an exposed API key, an unchecked input, a missing auth check. This prompt from hackSultan tells Claude to act as a security auditor and walk through the whole codebase looking for the standard OWASP top-10 issues. Returns a clean punch-list: severity, file, fix. https://escape9to5.ai /armory/audit-qa/visual-verify-playwright https://escape9to5.ai /armory/audit-qa/visual-verify-playwright Thu, 23 Apr 2026 00:00:00 GMT Audit & QA Every UI change ends with a screenshot. If it doesn't look right, it didn't ship right. TypeScript + build passing does NOT mean your UI looks right. A page can compile cleanly and still render broken — wrong layout, overlapping elements, missing assets. The only way to catch this before shipping is to actually LOOK at it. https://escape9to5.ai /armory/agent-frameworks/everything-claude-code https://escape9to5.ai /armory/agent-frameworks/everything-claude-code Wed, 22 Apr 2026 00:00:00 GMT Agent Frameworks The other starter pack for Claude Code. Skills, memory, and habits out the box. everything-claude-code is a ready-made pack of skills, instincts, and habits that plug into Claude Code when you install it. You get memory that survives across sessions, security sweeps built in, and a research-first habit that stops your agent from guessing when it should look things up. https://escape9to5.ai /armory/agent-frameworks/hermes-agent https://escape9to5.ai /armory/agent-frameworks/hermes-agent Wed, 22 Apr 2026 00:00:00 GMT Agent Frameworks A personal AI agent that lives in your terminal. Memory + 40 tools out the box. Hermes is a standalone agent — you install it and it runs in your terminal on its own, separate from Claude Code. It ships with memory that carries across sessions, plus 40+ tools already wired up (browser, files, scripts, SMS, Telegram). https://escape9to5.ai /armory/agent-frameworks/karpathy-skills https://escape9to5.ai /armory/agent-frameworks/karpathy-skills Wed, 22 Apr 2026 00:00:00 GMT Agent Frameworks Karpathy's coding advice, packaged as a CLAUDE.md your agent actually reads. Andrej Karpathy (the guy who taught the world how LLMs actually work) has posted years of notes on where AI coders go wrong — overengineering, missing obvious bugs, skipping tests. Someone packaged all of it into a single CLAUDE.md file you drop into any Claude Code project. https://escape9to5.ai /armory/agent-frameworks/superpowers https://escape9to5.ai /armory/agent-frameworks/superpowers Wed, 22 Apr 2026 00:00:00 GMT Agent Frameworks The one Claude Code install everyone should make first. Superpowers turns Claude Code from a chat loop into a disciplined builder. Install it once and every new project gets the same baseline: brainstorm before code, write a plan before touching files, verify it works before claiming done. https://escape9to5.ai /armory/agentic-autonomy/browser-harness https://escape9to5.ai /armory/agentic-autonomy/browser-harness Wed, 22 Apr 2026 00:00:00 GMT Agentic & Browser The fastest way to let Claude drive your browser. Self-healing, works on any site. Need Claude to log into something, scrape a page behind a paywall, fill out a form, or click through a real app? browser-harness spins up a real Chrome that Claude can control — screenshots, clicks, typing, iframes, popups, and it restarts itself if anything breaks. https://escape9to5.ai /armory/agentic-autonomy/chrome-devtools-mcp https://escape9to5.ai /armory/agentic-autonomy/chrome-devtools-mcp Wed, 22 Apr 2026 00:00:00 GMT Agentic & Browser Give your agent the same debug tools you use in Chrome's inspector. The Chrome DevTools panel — the thing you see when you right-click and hit Inspect — has network tabs, console logs, performance profilers, and memory tools. This plug exposes all of that to Claude. https://escape9to5.ai /armory/agentic-autonomy/cursor-cli https://escape9to5.ai /armory/agentic-autonomy/cursor-cli Wed, 22 Apr 2026 00:00:00 GMT Agentic & Browser Cursor's AI coding agent, in your terminal — no editor required. If you're already paying for Cursor and don't want a second AI subscription for Claude Code, Cursor CLI gets you the 'agent in a terminal' experience on your existing plan. Multi-file edits, a `/debug` for tricky bugs, customizable slash commands. https://escape9to5.ai /armory/agentic-autonomy/karpathy-autoresearch https://escape9to5.ai /armory/agentic-autonomy/karpathy-autoresearch Wed, 22 Apr 2026 00:00:00 GMT Agentic & Browser Karpathy's proof-of-concept for letting AI agents do research on their own. Andrej Karpathy built this to show what it looks like when you point AI agents at a scientific problem and let them iterate — they propose experiments, run them, read the results, propose new ones. Minimal babysitting. https://escape9to5.ai /armory/agentic-autonomy/playwright https://escape9to5.ai /armory/agentic-autonomy/playwright Wed, 22 Apr 2026 00:00:00 GMT Agentic & Browser Claude can open a browser, click, type, and screenshot. Works out the box. Playwright is Microsoft's browser automation toolkit. The Claude Code integration lets your agent open a page, click, type, take screenshots, check the console, and handle popups — all with no setup. https://escape9to5.ai /armory/business-marketing/larry https://escape9to5.ai /armory/business-marketing/larry Wed, 22 Apr 2026 00:00:00 GMT Business & Marketing Loya's own TikTok slideshow automation. 1M+ views, $670/mo MRR, running on an old gaming PC. Full pipeline: researches competitors, generates slide images, adds text overlays, posts via Postiz to TikTok + cross-posts to IG/YouTube/Threads, tracks analytics, adjusts hooks and CTAs based on what converts. https://escape9to5.ai /armory/business-marketing/marketing-skills https://escape9to5.ai /armory/business-marketing/marketing-skills Wed, 22 Apr 2026 00:00:00 GMT Business & Marketing Claude Code skill pack for marketing. SEO, CRO, ads, copy, analytics — all as slash commands. marketingskills is a plug-in pack of marketing workflows for Claude Code. Install it and you get slash commands like /directory-submissions (auto-submit your product to Product Hunt + G2 + AI directories), /competitor-profiling (deep competitive research), /copywriting audits, /SEO analysis, and more. https://escape9to5.ai /armory/business-marketing/postiz https://escape9to5.ai /armory/business-marketing/postiz Wed, 22 Apr 2026 00:00:00 GMT Business & Marketing Schedule and post to every social platform from one API. Works with Claude Code. Postiz lets you post + schedule to Instagram, TikTok, YouTube, X, Threads, LinkedIn — all from one hosted API or a self-hosted Docker instance. Has a Claude Code skill already written (the postiz skill we use daily). https://escape9to5.ai /armory/business-marketing/ui-ux-pro-max https://escape9to5.ai /armory/business-marketing/ui-ux-pro-max Wed, 22 Apr 2026 00:00:00 GMT Business & Marketing The skill that turns generic AI-generated UI into something that looks paid-for. Install this Claude Code skill and run `/ui-ux-pro-max` after any design change. It audits your UI for the little things that make AI output feel cheap — inconsistent spacing, weird hover states, color clashes, missing loading states — and fixes them. https://escape9to5.ai /armory/content-visual/gpt-image-2 https://escape9to5.ai /armory/content-visual/gpt-image-2 Wed, 22 Apr 2026 00:00:00 GMT Content & Visuals OpenAI's new image model. Better at photorealism + brand work than almost anything else. GPT-Image-2 is OpenAI's upgrade over their last image model. Creators are posting side-by-sides where one restaurant photo turns into a full brand guideline, and head-to-head tests where it beats Nano Banana Pro on the same prompt. https://escape9to5.ai /armory/content-visual/hyperframes https://escape9to5.ai /armory/content-visual/hyperframes Wed, 22 Apr 2026 00:00:00 GMT Content & Visuals Write HTML. Render video. The agent-native alternative to Remotion. Hyperframes lets you build videos the same way you build a webpage — write HTML, style with CSS, animate with GSAP or Three.js, then render to MP4. Built by HeyGen for Claude Code to drive directly (ships with skill packs, CLI lint + preview + render). https://escape9to5.ai /armory/content-visual/nano-banana-pro https://escape9to5.ai /armory/content-visual/nano-banana-pro Wed, 22 Apr 2026 00:00:00 GMT Content & Visuals Google's image model with the silly name. Does text-to-image + image-edit at 1K/2K/4K. Nano Banana Pro is Google's Gemini 3 Pro Image model, accessed through the kie.ai API. It handles clean text-to-image, photo-style edits (change one thing, keep everything else), and renders up to 4K. Good at faces, brand graphics, and photo retouching. https://escape9to5.ai /armory/content-visual/open-generative-ai https://escape9to5.ai /armory/content-visual/open-generative-ai Wed, 22 Apr 2026 00:00:00 GMT Content & Visuals Self-hosted alternative to Higgsfield with 200+ models. If you have the GPU, no more subscriptions. Open-Generative-AI packages 200+ image and video models into a single self-hosted studio — Flux for images, Kling and Veo for video, lip sync, the whole lineup Higgsfield charges you for. MIT licensed, open source.