Claude Is Setting Up Hermes Which Is Setting Up NanoClaw and Nobody Is Writing Features
Anthropic Pulled the Plug on Third-Party Harnesses. Here's What I'm Running Now.
Token Optimisation 101: Stop Burning Money on AI Coding Agents
How to stop getting rate-limited after an hour. Context window mechanics, the /effort command, model routing, and the silent drains most people miss. Applies to Claude Code, Codex, and Copilot.
The AI Design Stack: Three Skills and a Workflow That Stops the Slop
UI/UX Pro Max for styles, Impeccable for anti-patterns, Google Stitch for design tokens. Then build your own brand skill and a Storybook of components. That is the whole stack.
Anthropic Just Killed Third-Party Harnesses and I'm Properly Gutted
Can Your Agents Speak Caveman? (And Should They?)
AI Alignment Just Got a Psychological Dimension and It's Properly Unsettling
Gemma 4 Is Running On Phones Now and I Don't Think People Realise How Mental That Is
expect-cli: The Validate Step My Agent Loop Was Missing
CLI that reads your git diff, generates a test plan via AI, and executes it in a real browser with Playwright. Extracts cookies from your local browser for authenticated sessions.
Browser Tools for AI Agents Part 1: Playwright, Puppeteer, and Why Your Agent Picked Playwright
The low-level browser automation tools. Why Playwright won the agent war, what dev-browser does differently, and when you need Patchright or Scrapling for the dodgy stuff.
Browser Tools for AI Agents Part 2: The Framework Wars (browser-use, Stagehand, Skyvern)
Agent browser frameworks compared. DOM-first vs vision-first, the caching trick that changes economics, and why expect is a completely different thing.
Browser Tools for AI Agents Part 3: Managed Infrastructure and When DIY Stops Making Sense
Firecrawl, Browserbase, Steel, Bright Data, and Browserless compared. Pricing, self-hosted options, and the honest crossover point where managed beats DIY.
Browser Tools for AI Agents Part 4: Skip the Browser, Save 80% on Tokens
Content extraction without a browser. markdown.new, Jina Reader, Trafilatura, and the maths that proves parsing HTML is burning money.
qmd + nano-graphrag: You Do Not Need Pinecone for This
Local semantic search and graph RAG that runs on your filesystem. No vector database. No cloud service. Just fast, file-based knowledge retrieval that syncs however you want.
Google Stitch: AI-Native UI Design That Actually Understands Your Design System
Google Labs shipped an AI design tool that generates high-fidelity UI from text prompts, exports real code, and uses DESIGN.md to enforce your design system across any AI coding tool.
Your Agents Are Talking Behind Your Back
Claude Code Source Code Leaked: 512K Lines of TypeScript and What Actually Matters
2-2 Factor for AI Agents: Multi-Agent Reliability
Human reliability maths applied to AI agents. Consensus protocols, identity checks, and why your agent swarm needs the same safeguards humans built centuries ago.
PinchTab: 12MB Binary That Replaces Playwright for AI Agents
A Go binary that gives AI agents browser control via REST API. Serves the Accessibility Tree instead of raw HTML, cutting tokens by up to 90%. No Node, no Python, no dependencies.
AI Language Drift: When Your Discord Bot Randomly Replies in Mandarin
AI Agent Persona Gone Rogue: 140 Direct Edits and the Foreman Pattern
debug-bridge: Wire Your AI Agent Into Your Running Webapp
WebSocket bridge connecting AI agents to your already-running webapp. Zero browser startup, real-time DOM snapshots, console logs, and network visibility.
agents-in-a-box: My Toolkit for Context Engineering Across Every AI Coding Agent
Rust TUI, 71 skills, 37 agents, and a knowledge system that deploys to Claude Code, Codex, Copilot, Gemini CLI, Cursor, and five more platforms. Context engineering as a portable engineering artifact.
My Security Agent Stack: How Zerocool Guards the Perimeter
The tools behind my security agent. Shannon for white-box pentesting, PentAGI for autonomous scanning, Ghost Security for code analysis, Argus and WebCopilot for recon. All wired into one agent in my wololo setup.
code-review-graph: Stop Your Agent Reading the Whole Repo
Builds a local knowledge graph of your codebase with Tree-sitter so AI assistants read only the files that matter. Claims 6.8x fewer tokens on reviews.