Full-text search across 61579 articles. Combine with topic and date filters; results sorted by relevance.
A developer built a 'bug-reporter' Skill for Claude Code that automatically turns test failures into structured bug reports with evidence-based reproduction steps. The Skill uses guardrails to prevent fabricated reports …
Context engineering, not larger context windows, is the key to efficient AI-assisted coding, argues a developer in a technical thread. Over-providing context leads to higher latency, increased cost, and diluted focus due…
Anthropic removed over 80% of the Claude Code system prompt for its newest models Opus 5 and Fable 5, sharing lessons learned about writing system prompts, skills, and Claude.MDs for them. A user reported that Opus 5 is …
Claude Code proposes a shift from simple cost prediction to provable bounds in AI-driven query optimization, using quantile regression or conformal prediction to output confidence intervals instead of single-point estima…
A new Claude Code skill called Wrap, created by developer ravsau, lets users check whether a session is safe to close by running the /wrap command, which reviews the current conversation and returns either a STOP verdict…
Masih Moafi, developer of the terminal coding agent Elpis, discovered that AI agents he used to build the project introduced bugs and misleading commit messages. One agent reverted a core context-saving feature in a buri…
Anthropic's Claude Code agentic loop consumes far more tokens than its current pricing reflects, creating a deficit that favors developers, according to a technical analysis. The analysis warns that the current cost-to-p…
A technical governance blueprint for 2026 outlines that security approval for AI coding tools must rely on repository-level evidence, including a live inventory of models, SDKs, and MCP servers, enforced policy at the pu…
Rabbitty, a native Mac terminal for running AI coding agents in parallel, released Beta 0.1.0-beta.53 with faster integration panels, background prefetching for pull requests and Jira, and a fix for a Dock-launch race co…
Anthropic announced that Claude Code on desktop now works with the iOS simulator, allowing developers to build and run iOS apps with the simulator opening in a panel next to the conversation. The feature is available tod…
A new API called Smoke Break offers coding agents a simulated break, returning a brand, cough, remark, and break count via POST /smoke, with a GET /count endpoint for total breaks. The free API allows ten breaks per minu…
A developer built an MCP server for Claude that enforces hydration checks via webcam, using YOLOv8 nano to detect a cup or bottle every 30 minutes. If the state is 'locked,' Claude refuses to help with coding. The system…
A Hacker News user asks whether OpenCode and Kimi K3 are better than Claude Code, with a commenter noting that it depends on the criteria, specifically that if cost is the primary consideration, then yes, they are better…
Anthropic's Claude Code now includes a screen reader mode, activated by running `claude --ax-screen-reader`, which replaces the visual terminal UI with plain, linear text compatible with screen readers such as VoiceOver …
A developer created a /council skill for Claude Code that convenes a five-advisor council to pressure-test decisions, plans, or ideas. Based on Andrej Karpathy's LLM Council, the skill uses five subagents with distinct p…
A developer shared global CLI aliases that launch coding agents (agy, codex, claude, opencode) inside timestamped tmux sessions, bypassing permission prompts for faster iteration.
In December 2025, Cline, an AI coding assistant with over 5 million users, suffered a four-hop supply chain compromise after a misconfigured GitHub issue triage workflow allowed any user to inject malicious commands via …
A developer open-sourced a macro execution layer for coding agents that reduces the number of model turns by executing routine workflow steps together. On a 60-task DeepSWE benchmark, the macro approach with backward rea…
AgentGrove, an open-source local developer workspace for AI coding agents built with a Rust backend and SolidJS frontend, has been released. The tool supports cross-platform use on Linux, macOS, and Windows, and features…
A developer built Cortex, a local memory system for coding agents that stores session context in a single SQLite file. Cortex uses deduplication, typed provenance links, and a ranking engine to pack relevant recall into …