Topic · Concept
Coding agents
Coding agents are AI systems that take a software task, such as fixing a bug or adding a feature, and carry it out over many steps: reading code, editing files, running commands and tests, and returning a change for review. This page covers how they work, how they are measured and where they fail.
Fact file
Canonical agent vs workflow definitionAnthropic, "Building effective agents," December 19, 2024
SWE-bench VerifiedOpenAI stopped evaluating on it (Feb 23, 2026), citing flawed tasks and contamination; recommends SWE-bench Pro
Claude Code default modeAuto mode is the built-in starting permission mode for interactive terminal and VS Code sessions from v2.1.283
Example productsClaude Code (GA May 22, 2025), Codex (cloud agent May 16, 2025), Cursor, GitHub Copilot
Shared instruction formatAGENTS.md, a founding project of the Agentic AI Foundation (Dec 9, 2025)
Start here
Latest on Coding agents
Oct 9Agents · NewsThe Week in AI #2: Sonnet 5.5, OpenAI DevDay and a gated Gemini 4Oct 9Agents · NewsThe Week in AI #1: Opus 5.5, GPT-6 Sol and Grok 4.7 land within 48 hoursOct 9Developer tools · NewsOpenAI's Agents API puts the Codex harness behind one endpoint: what developers get, and what they give upOct 9Developer tools · NewsCursor's latest release puts your local agents on your iPhoneOct 9Developer tools · NewsGitHub Copilot can now drive your desktop apps: what computer use does and how to turn it onOct 9Agents · AnalysisThe case against unattended agents: what went wrong, and how sandboxing actually worksOct 9Agents · BenchmarksSWE-bench, Terminal-Bench and METR: what coding-agent evaluations measure, and what their results don't proveOct 9Agents · CompareClaude Code vs Codex vs Cursor: features, prices and limits compared (October 2026)Oct 9Apps · ExplainerChatGPT agent mode, explained: what it was, why it's gone, and what ChatGPT Work does insteadOct 9Agents · ExplainerOpenAI Codex explained: where it runs, which models and limits you get, and how its sandbox works
Nothing in this format on this page — try another filter or the next page.
About this hub
Coding agents are AI systems that take a software task, such as fixing a bug or adding a feature, and carry it out…