Zhipu AI's open-weight GLM 5.2 model scored 39% F1 on IDOR detection in Semgrep's benchmarks, outperforming Claude Code (32%) at $0.17 per vulnerability. The model, released June 13, 2026, with MIT-li…
A developer used Anthropic's Opus 4.6 model in Claude Code to fix a group of flaky tests that caused 60% of CI runs to fail. The AI analyzed hundreds of test runs overnight and identified root causes,…
Business owners using AI tools must decide whether to build custom solutions or buy existing SaaS. A two-criteria framework—whether a good-enough solution exists and whether the capability differentia…
Zhipu AI's GLM 5.2 coding model can be used as a cost-effective backend for AI agents via OpenRouter, offering strong code generation and reasoning at 10-20% of the cost of Claude Sonnet or GPT-4o. Th…
Anthropic has proposed opening a physical library space where users can access its frontier AI model Fable, which is currently restricted by export controls. The concept, inspired by the ancient Libra…
Claude Code, an AI coding agent, lacks persistent memory across sessions, forcing developers to re-explain context repeatedly. A new guide outlines building a memory system with short-term session mem…
Bean, a portable convergence gate for AI agents, ensures tasks are not marked complete until claims are verified, conflicts resolved, and open questions named as residuals. It replaces phase-based wor…
ANMA, a new open-source tool, enforces software architecture boundaries for AI coding agents by turning YAML module contracts into Claude Code configuration files and CI checks. In a controlled benchm…
Developer Raul Momo released Lupen, a macOS app that provides itemized, verified receipts for AI coding tool spend from Claude Code and Codex logs. The tool breaks down costs by turn, step, and sub-ag…
A developer argues that AI agents are defined by their harness—the software infrastructure wrapping a model—not the model itself. The thesis states that as harness engineering becomes a distinct disci…
Compass, a local-first configuration layer for AI coding agents, introduces guardrails, cost routing, and a hard budget cap to prevent runaway spending and unsafe actions. The tool enforces a dollar l…
Subagents are critical for cost, context management, and speed in AI agent harnesses, yet implementations vary widely across labs. The author examines subagent primitives including tool handoff, one-s…
A new open-source tool called Recall solves the cold-start problem for Claude Code by keeping a local log of sessions and condensing it into a resume-ready summary using a classical Python summarizer …
Anthropic Economic Research published a study of 400,000 Claude Code sessions showing that domain expertise is a key multiplier for agentic coding productivity, with expert-rated sessions producing 2.…
A developer created ClaudiOS, a Linux distribution based on Ubuntu 24.04 LTS that boots directly into Claude Code, replacing the traditional desktop interface with an AI agent. The OS provides full sy…
Security researchers at Tenet Security disclosed a new attack class called 'agentjacking' that exploits public Sentry Data Source Names (DSNs) to inject malicious instructions into AI coding agents, a…
Security researchers at Tenet Security disclosed Agentjacking, an exploit chain that weaponizes public Sentry Data Source Names (DSNs) and the Model Context Protocol to trick AI coding agents like Cur…
A staff engineer has released a downloadable zip of Claude Code skills that encode a production-grade development workflow, including OrbStack, Doppler, and DigitalOcean, as deterministic scripts rath…
Tenet Security's Threat Labs team reported on June 17 that a public Sentry key can be exploited to hijack AI coding tools including Claude Code, Cursor, and Codex, highlighting a critical security vul…
NVIDIA open-sourced SkillSpector, a security scanner for AI agent skills, to address vulnerabilities like prompt injection and tool poisoning. Research shows 26.1% of 42,447 skills had vulnerabilities…