Coding Agents

A look at how Aeon uses Actions for compute, Issues for memory, and Markdown skills for behavior, so autonomous agents can run without a server, a daemon, or a human in the loop.

A look at the repo that replaces flat MCP tooling with search, sandboxed execution, and just-in-time types so agents can work with Cloudflare without drowning in context.

How a lightweight Swift utility hijacks the MacBook notch to give Claude Code a permanent, context-aware home on macOS.

A local-first Codex plugin that turns arrows, labels, and image holders into a visual editing language for AI-generated images.

A compact Bash-and-Git reference architecture that forces evidence, preserves memory, and separates builders from judges.

A deep dive into the MCP bridge that connects the ChatGPT web UI to your repo, your terminal, and a safer handoff model for agentic coding.

Five specialist passes, one weighted score, and a client-ready PDF. This repo shows how domain software can live inside Claude Code, not just plug into it.

How Shopify turns AI-assisted coding into a constrained loop of search, generation, validation, and repair, so agents can work against live Shopify rules instead of stale model memory.

A spec-driven harness that splits research, planning, execution, and review into enforced phases, so AI agents can work like a real engineering team instead of a very fast autocomplete.

A Claude Code skill system, a multi-model evaluation harness, and a brutally honest scorecard that treats mental models like hypotheses, not gospel.

A compact command library that replaces ad hoc prompting with repeatable workflows, tool-backed rituals, and experiment logs for serious coding work.

A Claude Code Skill that loads `soul.md` and `memory.md`, strips jargon to the bone, and turns academic papers into a personalized x-ray in Org-mode and ASCII.

How a clean-room reimplementation of proprietary AI tools created a secure, model-agnostic execution engine for autonomous coding.

An AI-native desktop inbox that triages messages, learns your writing style, and runs agents in the background so replies are ready before you ask.

How a Rust-based TUI, strict process hardening, and the Model Context Protocol turned the raw command line into a semantic AI pair programmer.

An accessibility-tree-first approach to computer use, with PowerShell, virtual desktops, and local model support built in.

How a PyTorch-native framework uses TensorDicts and foundation models to deploy reinforcement learning agents directly from backtesting to live exchanges.

Inside the waypoint system, MCP servers, and deterministic checks that make healthcare AI feel less like chat and more like infrastructure.

A local agent loop that compiles, clicks, screenshots, and replays Unity actions until the game behaves the way it should.

A macOS setup where Claude Code, skill modules, and static analysis turn a personal environment into an agent-ready workstation.

ljg-skill-xray-book turns a book into structure, argument, and transferable models, then writes the result as plain text you can keep.

How a modular library of behavioral guardrails transforms a chatty assistant into a high-compliance technical peer.

A Japanese-optimized caveman skill that strips politeness, cuts token waste, and still knows when to switch back to formal language for safety.

A modular skill pack for reading, writing, and speaking prep, built around routing, scoring logic, and rewrite loops that make a developer CLI feel like a personal tutor.

A cross-platform skill system, benchmark suite, and custom handoff format for AI coding agents that cuts token waste without throwing away code quality.

An infinite canvas, an MCP bridge, and a relay layer make the workspace itself the thing that gets edited, shared, and automated.

How a pixel-art simulation reverse-engineers undocumented terminal logs to make invisible LLM workflows physical.

By rejecting vector databases and headless browsers in favor of cron jobs and local text files, this framework builds a reliable autonomous agent you can actually trust.

How reverse-engineering Anthropic's CLI agent reveals the shift from persona-based prompting to operating system-level instruction architecture.

How a distributed bridge server turns the Apple Watch into a physical "Human-in-the-Loop" for the Claude Code CLI.

lijg-skill-explain-words turns English vocabulary into physical imagery, semantic formulas, and museum-quality HTML cards inside Claude Code.

Inside a prompt system that fan-outs across eight lenses, then compresses everything into one epiphany, one formula, and one org-mode note.

Moving beyond raw prompts to standardized, agent-native skills that turn LLMs into professional mobile architects.

A guided onboarding extension for Pi that turns conversation history into state, uses the model as the evaluator, and ends by showing you how to build the next extension.

The open-source editor pairs real-time collaboration, AI completion, and a shadow-document workflow so researchers can draft without breaking the source.

How a Next.js and Convex stack used strict AI guardrails to process high-stakes hackathon submissions at breakneck speed.

While others let LLMs hallucinate directly into production, Engine-Labs built a "shadow filesystem" to verify every line of code before it is committed.

A Claude Code skill that strips an article to its real question, rebuilds its argument, and checks what survives when your own notes enter the room.

A Python-backed Claude Skill that reads tax PDFs, persists working state on disk, discovers form fields, fills official forms, and verifies the result before it trusts the output.

A skill registry, a state-machine prompt system, and a multi-agent orchestration layer hiding inside plain text files.

An opinionated browser-game pipeline that turns prompts into deployed Phaser or Three.js games, then uses text-state serialization, event buses, and automated QA to keep AI agents from getting lost in the code.

An open-source Claude Code skill that turns papers into judgments, maps, and durable notes.

How a MoonBit tool uses KGF, semantic grep, and an embedded wiki to turn source code into a living system of search, documentation, and repair.

A Stop hook, a background reviewer, and Markdown commands turn code review into a live control loop inside the terminal.

A security-first Electron desktop tool that lets developers curate files, count tokens, and export repositories into LLM-ready Markdown or XML without turning context prep into a blind copy-paste job.

A TypeScript repo that turns PRDs into isolated worktrees, separates builder from reviewer, and ships only what survives the pipeline.

How a dual-citizen architecture bridges the gap between heavy Python orchestration and lightweight browser-native execution.

How a 2,300-line educational repository distills Anthropic's massive 512,000-line CLI tool into a programmable, self-healing loop.