Tag
A tweet sharing best practices for AGENTS.md optimization and dead code cleanup, with ready-to-copy prompts. It covers use cases of cleaning up 800K+ lines of dead code with GPT-6 Astra + Codex, and optimizing CI duration with GPT-6.1 Sol.
Matt Pocock shares a 'prompt of the day' for cleaning up a messy AGENTS.md: remove no-ops, use progressive disclosure, and move instructions to CODING_STANDARDS.md, delegating the work across three increasingly radical subagents into a single PR. He argues AGENTS.md files are the biggest source of token inefficiency in agent workflows.
A practitioner reports that Karpathy's ASD-STE100 controlled-language tip genuinely improves LLM response clarity (including HTML artifacts) when applied selectively, and shares a concrete self-improvement prompt that samples recent transcripts to derive custom clarity rules for the user's AGENTS.md file.
The article explains that Claude Code only reads the AGENTS.md file when telemetry is enabled due to a server-side feature flag, causing silent failures for users with telemetry off, and proposes a workaround using CLAUDE.md imports.
Claude Code has been updated to read AGENTS.md files when Claude.md is not present, improving its configuration handling.
Claude Code has been updated to version 2.1.277, adding support for AGENTS.md as a fallback configuration file when CLAUDE.md is not present in a folder.
Claude Code now supports AGENTS.md as a fallback configuration file when CLAUDE.md is not present, starting from version 2.1.277.
This article discusses the overengineering problem in Codex Sol AI models, which generate excessive code for simple tasks, and proposes an AGENTS.md approach to mitigate this issue.
Kent C. Dodds discusses a research paper showing AGENTS.md often doesn't improve task success while adding over 20% inference cost, and asks what should actually be placed in AGENTS.md, noting Claude Code reduced its system prompt by 80%.
A developer shares a comprehensive Agents.md template with core principles for building and maintaining trading systems using AI agents, covering strategy development, engineering standards, and scope discipline.
AGENTS.md summarized by Vercel engineers can save you 80% of tokens in project development. The core is to reduce AI's ineffective iterations and token consumption through engineering decisions and solution design.
Discusses how developers using OpenClaw can keep context for AI coding agents by using a handoff document (AGENTS.md) to track goals, files, failures, and decisions in messy coding sessions.
GitHub Copilot code review now supports AGENTS.md files, allowing users to customize for more context-aware reviews.
This article discusses the design concept of how Markdown files (such as AGENTS.md and SKILL.md) in Coding Agents effectively influence agent behavior through the Harness mechanism, emphasizing the importance of loading different contexts at the right time.
This paper evaluates whether repository-level context files like AGENTS.md or CLAUDE.md improve coding agent performance, finding that LLM-generated context files offer little benefit and may reduce efficiency, while developer-written files are better but still not clearly advantageous.
Google announces the official launch of Managed Agents in the Gemini API, enabling agent development with zero infrastructure overhead. The article also highlights AGENTS.md, an open format for providing context to AI coding agents, adopted by over 60k open-source projects.
A methodology for autonomously training transformer language models on a single consumer GPU, structured in six stages with verification gates and AGENTS.md specs for orchestration frameworks like OpenClaw.
A tweet explains how to clone Andrej Karpathy's AGENTS.md workflow into Codex by pasting the 65-line file into Global Custom Instructions, making Codex behave more like a structured autonomous coding agent.
A quick tip showing how to copy Andrej Karpathy's 65-line Agents.md config into Codex Global Custom Instructions for an effective minimalist setup.
Discusses how to use a concise global Codex prompt for Claude Code to avoid over-cautiousness, citing Karpathy's 65-line Agents.md configuration as an example.