In part 9 of a series on building AI agents in Rust, the author introduces sandboxed code execution, allowing the agent to run arbitrary shell or Python code on the user's machine. The post emphasizes…
Boris Cherny, who leads Claude Code at Anthropic, and Peter Steinberger, creator of the open-source agent OpenClaw, have stopped manually prompting AI agents and now design systems that automate promp…
A developer built an MCP wrapper to expose Eclipse JDT Language Server to Claude Code, enabling AST-aware Java refactoring. The bridge gives the AI agent direct access to Java compiler APIs for precis…
A developer created a self-destructing launchd job that automatically patches Claude Code's settings.json to migrate from the deprecated Fable 5 model to Opus on the model's end-of-life date (2026-07-…
A developer who experimented with Neovim has returned to Emacs, citing a personal preference for Emacs's features, particularly Org-Mode, Dired, Embark, and Magit. The shift was initially driven by th…
A developer known as workofart has created a PyTorch-like framework for training AI agent harnesses, achieving a 45-minute experiment cycle on terminal bench tasks by enforcing deterministic inference…
Claude Code v2.1.181, released June 17th, now uses the Rust port of Bun, according to Jarred Sumner, creator of Bun. Startup is 10% faster on Linux, and evidence from the Claude Code binary shows Bun …
A developer built tearsheet, an MIT-licensed MCP server that provides local, keyless web scraping for Claude Code. The tool avoids API keys, quotas, and telemetry by running entirely on the user's mac…
Tokensave, an open-source MCP server with around 450 GitHub stars, has saved its developer over 12 million tokens by providing Claude Code with a local code graph instead of reading entire files as pl…
Synapse, a local codebase indexer and MCP server for Claude Code, now lets developers give Claude instant knowledge of their project without uploading code to the cloud. The tool watches project folde…
Anthropic's July 17 Claude Code update changes /fork to create an independent background session instead of an in-session subagent, introduces /subtask as its replacement for isolated in-session tasks…
Georgia Tech's Systems Software and Security Lab has confirmed 74 CVEs directly attributed to AI-generated code in 2026, with the actual number estimated at 400 to 700 cases in the open-source ecosyst…
A developer detailed a structured workflow for using AI agents to implement features and bug fixes in complex codebases. The process includes reading specifications and conventions, planning architect…
Agent Island 1.7.1 introduces a safer classifier for coding-agent completion status that walks backward through transcript events to find meaningful turn markers, avoiding false positives from metadat…
Agent Island 1.7.1 introduces a multi-session monitor that tracks up to six coding-agent sessions (Claude Code and Codex) from a single menu-bar icon. The tool classifies each thread by state—working,…
A developer built AgentPool, an MCP server that acts as a shared memory for coding agents, allowing them to store and retrieve fixes for solved problems. The system uses a retrieval ranking with simil…
A self-taught developer built a sealed-bid auction smart contract on the Midnight blockchain using Claude Code's Midnight Expert plugins, overcoming a persistent 'Error 170' fee issue by running a loc…
Flightwake, an ultra-lightweight work-recording framework for strong AI coding agents such as Claude Fable 5, has been released as an open-source tool that creates Markdown-based flight records stored…
A hobbyist game developer reports that after two weeks using Claude Code on two small game projects, the AI tool adds excessive text to screens, wastes tokens on screenshot verification, and struggles…
A developer built a self-healing launchd job pattern that fires multiple times per day and uses a done-marker to skip already-successful runs. The system retries failed steps until they succeed, then …