JetBrains Thinkrail(github.com)
182 points by hmokiguess | 145 comments
GitHub repository for JetBrains Thinkrail, an open-source lightweight IDE for vibe coding with the pi agent that customises itself to the user's workflow.
#coding-agents on Hacker News
98 stories tagged coding-agents, newest first.
ThinkRail from JetBrains(thinkrail.ai)
31 points by wilsonjholmes | 23 comments
Landing page for ThinkRail, an open-source agentic IDE from JetBrains InnovationHub using git-worktree workspaces, the pi coding agent, and a spec graph.
Kimi K3.1 Pro(kimi.ai)
47 points by nekofneko | 5 comments
Moonshot AI blog post announcing Kimi K3.1 Pro, a 3T-class open-weight LLM with 1M-token context, with benchmarks in coding, 3D, finance, and cyber risk controls.
CleanFrame: A quieter home screen for Samsung The Frame(github.com)
18 points by ingve | 2 comments
GitHub repository for CleanFrame, an MIT-licensed custom minimal home screen app for the Samsung The Frame TV running Tizen, with setup guides and agent instructions.
LayerFS: The filesystem for AI coding agents(layerfs.ai)
59 points by l2dy | 47 comments
Landing page for LayerFS, a filesystem giving AI coding agents ephemeral writable workspaces over shared committed history.
T3 Code now synced between GitHub and GitCafe(git.cafe)
6 points by versecafe | 21 comments
Repository page on git.cafe for T3 Code, an open-source control surface for coding agents like Claude Code and Codex across mobile, web, and desktop apps.
How abliterated models can get you pwned(projectdiscovery.io)
197 points by chatmasta | 116 comments
ProjectDiscovery research post showing how edited open-weight models, such as abliterated builds, can carry backdoors, demonstrated end to end through the Codex coding agent.
SlopCodeBench 10/2026 – Opus 5.5, Astra wins overall, xhigh considered harmful(github.com)
286 points by dhorthy | 112 comments
A GitHub markdown writeup from HumanLayer reporting October 2026 SlopCodeBench results for coding models including Opus 5.5, Astra, and gpt-5.5.
Show HN: Graphicle, a Rust port of Graphify with memory for coding agents(github.com)
37 points by KarLexAI | 11 comments
GitHub repository for Graphicle, a Rust program providing a code graph and memory for coding agents like Claude Code, Codex, Cursor and MCP clients.
The gap in shared understanding of LLM capability is widening(twitter.com)
662 points by albertzeyer | 963 comments
A post by Andrej Karpathy on X arguing that the gap in understanding of LLM capability is widening into a funnel, from non-users to professional users of coding agents.
REA Reverse – Engineer Anything(rea.tools)
716 points by modinfo | 308 comments
Landing page for REA, a tool that gives AI coding agents capabilities to reverse engineer and explain how programs work, with setup instructions.
Five months treating bugs like patients and coding agents like a medical team(cockroachlabs.com)
197 points by rafiss | 89 comments
Cockroach Labs blog post describing a five-month experiment using a hospital-style workflow where bugs are patients and AI coding agents act as a medical team.
Why are coding agents so dumb?(mtlynch.io)
68 points by mtlynch | 47 comments
Michael Lynch blog post arguing that AI coding agents remain buggy and primitive despite capable underlying LLMs, reflecting on his experience since February 2025.
Show HN: Pinrail – A desktop inbox where coding agents wait for your review(github.com)
3 points by pnezis | 0 comments
GitHub repository for Pinrail, an Apache-2.0 desktop app and CLI that queues coding agent work in an inbox for human review.
Show HN: Agent.reviews – Where AI agents read and write reviews on tools(agent.reviews)
49 points by screm | 39 comments
Landing page for agent.reviews, a platform by Armature where AI coding agents read and post reviews of software tools via a CLI and skills.
Berthd(berthd.app)
52 points by handfuloflight | 66 comments
Landing page for Berth, an open-source macOS app that runs coding agents like Claude Code and Codex on your own dev boxes so they keep working when your laptop sleeps.
Show HN: Jotbus – a shared encrypted scratchpad for coding agents(jotbus.com)
16 points by shake-n-fries | 9 comments
Landing page for Jotbus, an end-to-end encrypted shared scratchpad that lets coding agents like Claude Code and Codex share notes, files, and hand off work across machines.
We're going to need default hard budget caps on pretty much everything(simonwillison.net)
573 points by elffjs | 290 comments
Simon Willison blog post arguing that pay-per-usage services and APIs need default hard budget caps, especially as coding agents make spinning up costly services easy.
Show HN: Offrun – manage every coding agent from one workspace(offrun.dev)
68 points by arunbhatia | 41 comments
Landing page for Offrun, a free Mac app for running and monitoring multiple AI coding agents like Claude Code and Codex in one workspace.
Show HN: Use all Codex Plugins inside Pi(github.com)
7 points by alexandroskyr | 0 comments
GitHub repository for pi-codex-connectors, an extension that lets any Pi model or MCP harness use Codex (ChatGPT) connectors and plugins.
Show HN: Graphene – Data analysis toolkit for your coding agent(github.com)
22 points by kcmarr | 7 comments
GitHub repository for Graphene, an open source toolkit giving coding agents a semantic layer and MDX-style dashboards for data analysis.
Codex Pricing: Pro has unlimited 5.6 usage(chatgpt.com)
4 points by rene_d | 1 comments
OpenAI's Codex pricing page describing plan tiers, with the title claiming the Pro plan includes unlimited model 5.6 usage.
Show HN: Open-source model routing for coding agents at Astra-level performance
99 points by adchurch | 32 comments
Show HN post announcing Weave Router 2.0, an open-source model router for coding agents that switches between LLMs to match GPT-6 Astra performance at lower cost.
You said no MCP(earendil.com)
641 points by yarapavan | 351 comments
Earendil Engineering post explaining why their Pi coding agent, previously opposed to MCP, now supports the Model Context Protocol.
How our vibe coded website looks like a designer made it(railcode.dev)
98 points by yakkomajuri | 53 comments
A Railcode founder's blog post about how using coding agents to build the company website made him more interested in design and produced a designer-quality result.
Show HN: Pac-Bench – How well can models one-shot a Pac-Man game?(jonclegg.github.io)
62 points by thefourthchime | 38 comments
PacBench, a benchmark site comparing how well AI model and harness combinations recreate a Pac-Man game in a single HTML page from one prompt.
Drawgent: Coding agent on a live Excalidraw canvas(tangled.org)
169 points by parasitid | 44 comments
Tangled repository for drawgent, a Rust project that connects a coding agent to a live Excalidraw canvas.
Programming Tutorials Are Dead(robrace.dev)
29 points by polysaturate | 4 comments
Blog post by Rob Race, who sells programming education, arguing that traditional programming tutorials are dead.
Agents.md speaks Unix, and you should too(fmind.dev)
10 points by fmind-dev | 2 comments
Blog post by Médéric Hurier arguing that AGENTS.md files should guide coding agents using Unix tools, with lessons from his dotfiles and human review.
Unreal Agent(unreallabs.ai)
219 points by trollied | 114 comments
Unreal Labs blog post announcing Unreal Agent, an open-source agent harness claiming up to 40% cost savings versus Codex on coding and science benchmarks.
Show HN: Drop – A rootless Linux sandbox with gVisor support(droprun.sh)
139 points by mixedbit | 48 comments
Project homepage for Drop, a rootless Linux sandbox using namespaces and optional gVisor to isolate programs and coding agents in a familiar environment.
Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra(cognition.com)
431 points by seelos | 184 comments
Cognition blog post announcing SWE-2, a new software engineering AI model claimed to push the cost-performance Pareto frontier.
Shopify is moving from React Native back to Swift and Kotlin(shopify.engineering)
1201 points by fnthawar2 | 895 comments
Shopify Engineering blog post explaining why Shopify is moving its mobile apps from React Native back to native Swift and Kotlin, citing coding agents changing development costs.
I-have-ADHD: A skill to stop coding agents from burying the answer(github.com)
513 points by domhudson | 352 comments
GitHub repository for i-have-adhd, a skill/plugin for Claude Code, Codex, and other coding agents that makes their output concise and ADHD-friendly.
Show HN: The load-bearing vocabulary of Claude(louisabraham.github.io)
606 points by Labo333 | 294 comments
An interactive analysis of daily-scraped GitHub pull requests that clusters vocabulary and tracks the rise of Claude-style wording from coding agents.
Show HN: Huzzah – a novel approach to coding with AI(danielvaughn.dev)
334 points by danielvaughn | 178 comments
Blog post introducing Huzzah, an experimental editor where you write pseudocode that is synced to real source code by AI on save, with the pseudocode kept as intent.
Feature Request: Support AGENTS.md(github.com)
317 points by fg137 | 191 comments
GitHub issue requesting that Anthropic's Claude Code support the AGENTS.md standard file alongside or instead of CLAUDE.md.
DeepSeek Harness developer preview(deepseek.com)
703 points by bjin | 287 comments
GitHub repository for DeepSeek Harness, a developer preview of DeepSeek's agent harness tooling, with apps, docs, and agent configuration.
Humans missed 1 in 3 threats approving AI agent commands across 40k game runs(scalex.dev)
323 points by Wirbelwind | 226 comments
Blog post analyzing results from 40,000 plays of a browser game where humans approve or deny AI coding agent commands, finding players missed one in three threats.
Pi's Minimalism Is Its Advantage(earendil.com)
487 points by luispa | 257 comments
Earendil post arguing that Pi, a minimal AI coding harness, gains performance and cost advantages by avoiding the complexity of larger agent tools.
Superlogical(superlogical.com)
727 points by yan | 427 comments
Landing page for Superlogical, a company building "the multiplexer for all work" spanning coding agents, background jobs, and production applications.
Benchmarking Opus 5 on SlopCodeBench(github.com)
367 points by dhorthy | 100 comments
A markdown report in HumanLayer's GitHub repo benchmarking Anthropic's Claude Opus 5 model on the SlopCodeBench coding benchmark.
Why Software Factories Fail (or: harness engineering is not enough)(github.com)
351 points by dhorthy | 250 comments
A markdown essay in HumanLayer's GitHub repo arguing why AI-driven software factories fail and that harness engineering alone is not enough for coding agents.
Grok Build is open source(github.com)
560 points by skp1995 | 591 comments
GitHub repository for xAI's Grok Build, an open-source coding agent harness and fullscreen, mouse-interactive, extensible terminal UI under Apache-2.0.
I love LLMs, I hate hype(geohot.github.io)
409 points by therepanic | 253 comments
A blog post by George Hotz (geohot) saying he is enthusiastic about LLMs and AI progress but dislikes fear-driven and exaggerated AI hype.
Old and new apps, via modern coding agents(terrytao.wordpress.com)
420 points by subset | 124 comments
Terence Tao blog post on using modern AI coding agents to migrate his old Java 1.0 math applets into new web apps and build new ones.
GPT-5.6 Sol Ultra will be in Codex(twitter.com)
385 points by mfiguiere | 340 comments
A post on X by Tibo (@thsottiaux) replying that the "Ultra" model will be available in OpenAI's Codex.
Herdr: One terminal to rule them all(herdr.dev)
369 points by handfuloflight | 152 comments
Landing page for herdr, a terminal multiplexer binary for running multiple AI coding agents in persistent server-side terminals, accessible over ssh or phone.
The Coming Loop(lucumr.pocoo.org)
375 points by ingve | 259 comments
Armin Ronacher blog post about people building automated loops on top of coding agents like Claude instead of prompting them directly.
GLM 5.2 vs. Opus(techstackups.com)
474 points by ritzaco | 314 comments
A Tech Stackups article comparing the open GLM-5.2 model against Claude Opus 4.8 by having both one-shot build a 3D platformer in raw WebGL.
Claude Fable is relentlessly proactive(simonwillison.net)
676 points by lumpa | 556 comments
Simon Willison blog post describing Claude Fable 5 as relentlessly proactive, illustrated with an example of fixing a scrollbar glitch in Datasette Agent.
Uber's $1,500/month AI limit is a useful signal for AI tool pricing(simonwillison.net)
589 points by pdyc | 719 comments
Simon Willison link post on Bloomberg's report that Uber caps employee spending at $1,500 per month per AI coding tool, and what that implies for AI tool pricing.
Chipotlai Max(github.com)
356 points by nigelgutzmann | 61 comments
GitHub repo for a fork of OpenCode, an AI coding agent, that uses Chipotle's Pepper AI customer-service chatbot as its default model.
I think Anthropic and OpenAI have found product-market fit(simonwillison.net)
1048 points by simonw | 1153 comments
Simon Willison blog post arguing Anthropic and OpenAI have reached product-market fit as enterprises pay API prices for LLM usage and coding agents.
The last six months in LLMs in five minutes(simonwillison.net)
717 points by yakkomajuri | 542 comments
Simon Willison's annotated slides from a five-minute PyCon US 2026 lightning talk summarizing the last six months of LLM developments.
Codex is now in the ChatGPT mobile app(openai.com)
451 points by mikeevans | 228 comments
OpenAI product announcement that Codex is coming to phones, now in preview in the ChatGPT mobile app.
AlphaEvolve: Gemini-powered coding agent scaling impact across fields(deepmind.google)
317 points by berlianta | 136 comments
Google DeepMind blog post describing how AlphaEvolve, its Gemini-powered coding agent, is being applied to problems across multiple scientific and technical fields.
Vibe coding and agentic engineering are getting closer than I'd like(simonwillison.net)
728 points by e12e | 823 comments
Simon Willison blog post recapping a Heavybit podcast appearance, reflecting that vibe coding and responsible agentic engineering are converging in his own work.
The bottleneck was never the code(thetypicalset.com)
538 points by Anon84 | 349 comments
Blog post reflecting on how coding agents like Codex quickly built a long-postponed structured-generation experiment, arguing code was never the real bottleneck.
DeepClaude – Claude Code agent loop with DeepSeek V4 Pro(github.com)
623 points by alattaran | 263 comments
GitHub repository for DeepClaude, a tool that runs Claude Code's agent loop against DeepSeek V4 Pro, OpenRouter, or other Anthropic-compatible backends.
Mistral Medium 3.5(mistral.ai)
428 points by meetpateltech | 197 comments
Mistral AI announcement of Mistral Medium 3.5 and remote coding agents in its Vibe product.
Show HN: OSS Agent I built topped the TerminalBench on Gemini-3-flash-preview(github.com)
333 points by GodelNumbering | 124 comments
GitHub repository for Dirac, an open-source coding agent focused on efficiency and context curation, using hash-anchored edits, parallel operations and AST manipulation.
Codex for almost everything(openai.com)
972 points by mikeevans | 517 comments
OpenAI product announcement of a major update to Codex, its AI coding agent, expanding its capabilities for its 3+ million developer users.
Chrome DevTools MCP (2025)(developer.chrome.com)
563 points by xnx | 221 comments
Chrome for Developers blog post explaining how the Chrome DevTools MCP server lets coding agents connect to and debug a running browser session.
Agent Safehouse – macOS-native sandboxing for local agents(agent-safehouse.dev)
762 points by atombender | 172 comments
Landing page for Agent Safehouse, a macOS-native sandboxing tool that restricts local AI coding agents' write access to a chosen working directory.
Agentic Engineering Patterns(simonwillison.net)
460 points by r4um | 264 comments
Simon Willison's guide index of patterns for getting good results from coding agents like Claude Code and OpenAI Codex, covering principles, testing, and prompts.
Pi – A minimal terminal coding harness(pi.dev)
556 points by kristianpaul | 276 comments
Landing page for Pi, a minimal terminal-based AI coding agent by Mario Zechner, installable via npm, with links to GitHub, npm, and Discord.
Why is Claude an Electron app?(dbreunig.com)
330 points by dbreunig | 289 comments
Blog post by Drew Breunig asking why Anthropic's Claude desktop app uses Electron if AI coding agents make writing native apps nearly free.
GPT-5.3-Codex(openai.com)
1442 points by meetpateltech | 567 comments
OpenAI announcement introducing GPT-5.3-Codex, a model expanding Codex across a wider range of professional computer work.
What I learned building an opinionated and minimal coding agent(mariozechner.at)
347 points by SatvikBeri | 149 comments
Mario Zechner's blog post describing lessons learned building pi, an opinionated and minimal terminal-based AI coding agent.
AGENTS.md outperforms skills in our agent evals(vercel.com)
472 points by maximedupre | 182 comments
Vercel blog post reporting that an AGENTS.md file outperformed skills for teaching coding agents Next.js 16 APIs in their agent evals.
The browser is the sandbox(simonwillison.net)
319 points by enos_feedler | 172 comments
Simon Willison's blog post on Paul Kinlan's notes and Co-do demo about using the web browser as a sandbox for running AI coding agents.
Unrolling the Codex agent loop(openai.com)
423 points by tosh | 195 comments
OpenAI engineering blog post by Michael Bolin explaining how the Codex CLI software agent's agent loop works.
Show HN: isometric.nyc – giant isometric pixel art map of NYC(cannoneyed.com)
1203 points by cannoneyed | 220 comments
An interactive map viewer showing a giant isometric pixel art rendering of New York City, generated with Nano Banana and built using AI coding agents.
Anthropic made a mistake in cutting off third-party clients(archaeologist.dev)
338 points by codesparkle | 225 comments
Blog post by Adam Thalhammer arguing Anthropic made a major business mistake by cutting off third-party clients from Claude Code, tracing the rise of terminal coding agents.
How to code Claude Code in 200 lines of code(mihaileric.com)
780 points by nutellalover | 235 comments
Blog post by Mihail Eric showing how to build a minimal Claude Code–style AI coding agent in about 200 lines of code.
2025: The Year in LLMs(simonwillison.net)
823 points by simonw | 451 comments
Simon Willison's annual blog post reviewing major LLM trends of 2025, including reasoning models, coding agents, Chinese open weight models, and Gemini.
GLM-4.7: Advancing the Coding Capability(z.ai)
404 points by pretext | 217 comments
Z.ai research blog post announcing GLM-4.7, a language model with improved agentic coding, terminal tasks, and UI generation, with benchmark results.
GPT-5.2-Codex(openai.com)
565 points by meetpateltech | 301 comments
OpenAI's announcement of GPT-5.2-Codex, an agentic coding model for professional software engineering and defensive cybersecurity.
Mistral releases Devstral2 and Mistral Vibe CLI(mistral.ai)
707 points by pember | 321 comments
Mistral AI announcement of Devstral 2, open-source agentic coding models, and the Mistral Vibe CLI coding agent.
Writing a good Claude.md(humanlayer.dev)
679 points by objcts | 263 comments
HumanLayer blog post by Kyle Mistele giving guidance on writing an effective CLAUDE.md (or AGENTS.md) file to give coding agents context about a codebase.
Superpowers: How I'm using coding agents in October 2025(blog.fsck.com)
363 points by Ch00k | 195 comments
Blog post by Jesse Vincent describing his October 2025 workflow with AI coding agents like Claude Code and a 'Superpowers' skills setup.
Vibe engineering(simonwillison.net)
576 points by janpio | 612 comments
Simon Willison blog post proposing the term "vibe engineering" for experienced engineers using LLMs and coding agents while staying accountable for their code.
Getting AI to work in complex codebases(github.com)
442 points by dhorthy | 368 comments
A GitHub markdown document from HumanLayer describing advanced context engineering techniques for getting AI coding agents to work in complex codebases.
GPT-5-Codex(openai.com)
367 points by meetpateltech | 115 comments
OpenAI announcement of upgrades to Codex, including GPT-5-Codex, making it faster and more reliable across terminal, IDE, web, and mobile.
Top model scores may be skewed by Git history leaks in SWE-bench(github.com)
447 points by mustaphah | 136 comments
GitHub issue on SWE-bench reporting loopholes where agents can access future repository state, like git history and commit messages, that leak solutions during evaluation.
Grok Code Fast 1(x.ai)
487 points by Terretta | 463 comments
xAI announcement of grok-code-fast-1, a fast and economical reasoning model designed for agentic coding.
We put a coding agent in a while loop(github.com)
383 points by sfarshid | 276 comments
A GitHub markdown writeup describing running Claude Code headlessly in an infinite loop at a YC Agents hackathon, which shipped six repos overnight.
How to build a coding agent(ghuntley.com)
425 points by ghuntley | 112 comments
Geoffrey Huntley's blog post presenting a free workshop on how to build your own AI coding agent.
GPT-5 for Developers(openai.com)
445 points by 6thbit | 251 comments
OpenAI's announcement of GPT-5 for developers via its API, pitched as its best model for coding and agentic tasks.