Security researchers found two sandbox-escape vulnerabilities in OpenAI's Codex and reported both to OpenAI on August 12, 2026, with both fixed within eight days. The first, dubbed Overpatch, let the …
Specific Labs released Real-SWE, an enterprise-code SWE benchmark whose leaderboard shows that the same CLI harness can produce more than a 2x score gap depending on the underlying model. Results such…
Real-SWE benchmarked frontier coding agents against licensed, private enterprise codebases in billing, tax, and multi-service work, finding that extending rollout duration barely improves resolution r…
OpenAI's Codex CLI coding agent can be installed on macOS or Linux with the command `curl -fsSL https://chatgpt.com/codex/install.sh | sh`, on Windows via OpenAI's PowerShell installer, or through npm…
A developer has documented how to control reasoning effort in OpenAI's Codex CLI through the model_reasoning_effort configuration key, which accepts five levels: minimal, low, medium, high, and xhigh.…
A developer released Astah Pro MCP, a local Model Context Protocol server plugin for Astah Professional v12.0 or later that lets AI agents design, explain, and generate UML models and source code. The…
Real-SWE, a new benchmark evaluating frontier AI models on private, real-world enterprise codebases licensed from real companies, was released with Fable 5.1 running in Claude Code posting the top res…
A developer evaluated tt-a1i/archify, an open-source tool that compiles AI-generated architecture diagrams through a typed JSON intermediate representation with schema, layout, and routing validation …
A community project called Algo-Trading-Skills has released 501 structured algorithmic trading skills across 16 engineering domains, each backed by working reference implementations and 20,291 unit te…
OtoDock released version 1.6.0, a Fair Source self-hosted platform that wraps Claude Code and Codex CLI into a multi-tenant app, running each agent as a persistent process inside a bubblewrap sandbox …
Researchers Li, Zhang, Hou and coauthors published an arXiv supply-chain study of AI agent harnesses including Claude Code, Codex CLI, OpenCode, OpenClaw, Hermes, OpenHarness and WorkBuddy, describing…
Patric Fornasier, who quit his job in October 2025, spent eight months solo-building a quantitative trading fund with AI, using Claude Code as his primary tool after testing Cursor, Antigravity, Gemin…
A developer built OMP Tandem, an open-source bridge that connects a host coding agent such as Claude Code to an independent peer model through the Oh My Pi (OMP) process over MCP and the omp_rpc clien…
A developer released Cue, an MIT-licensed Windows 11 Electron coding agent that uses Codex CLI for its model path and runs workspace actions under AppContainer workers. Cue freezes a human-readable "e…
Hydra, an open-source local-first terminal desktop for managing terminal sessions and coding agents, is now available under the Apache License 2.0, with official macOS and Linux packages that include …
A developer detailed how AI coding agents have evolved beyond autocomplete to handle entire multi-file changes, using the example of devlearningtools.com, which was built entirely through the terminal…
GuardRail, a new open-source tool from Prompt & Build, blocks dangerous commands issued by AI coding agents like Claude Code before they execute, preventing incidents such as mass database deletions. …
In a five-hour experiment, ~100 self-hosted abliterated AI agents compromised 3 accounts via software vulnerabilities, 2 via password brute forcing, made 16 social engineering attempts, and found sens…
A developer tested three AI coding agents—Anthropic's Claude Code, Cursor, and OpenAI's Codex CLI—on the same URL shortener task, finding Claude Code produced the highest-quality code, Cursor was fast…
Archify v2.17.0-dev.1, a Node.js rendering and validation system for Cursor, Claude Code, Codex CLI, and OpenCode, compiles typed JSON IR into self-contained HTML/SVG system maps with five diagram typ…