Full-text search across 60212 articles. Combine with topic and date filters; results sorted by relevance.
A developer published claude-polyglot, a multi-language Docker development container for running Anthropic's Claude Code CLI with DNS filtering enabled by default. The container restricts network access to api.anthropic.…
A follow-up experiment by developer Blandinium comparing three frontier cloud models against three large open-weight models via OpenRouter found the larger models far more reliable at code optimization, with the extracte…
A developer argues that AI-generated website prototypes need an explicit code handoff boundary once they take on real accounts, orders, or data responsibility. The piece recommends testing handoffs by having an uninvolve…
Cognition released SWE-2, a coding model post-trained with reinforcement learning from Moonshot AI's 2.8T-parameter Kimi K3, scoring 50.0% on FrontierCode 1.1 Main — within 1 point of Fable 5.1 at 64% lower cost. SWE-2 h…
CodePress launched an engineering platform that automates code review and deploys AI agents to handle feature implementation, testing, and review cycles. According to the product site, CodePress agents verify their own w…
A developer released AuditAI Scanner, an open-source tool that detects a common authorization flaw in AI-generated code where service-role keys bypass row-level security, letting one tenant read another's data. The scann…
A 30-day experiment found that AI coding tools handled CRUD operations, Zod schemas, and initial scaffolding but increased the developer's actual output by only about 20%, because writing syntax is roughly 20% of the job…
Slack announced Slackforce Surfaces, a feature that lets users describe a tool to Slackbot and have AI build interactive reports, polls, dashboards, presentations, and microsites inside chats, pulling data from conversat…
Google Cloud's threat intelligence blog warned that adversaries are actively exploiting prompt injection vulnerabilities against coding agents deployed in production environments. Google identified the attack pattern as …
DeepSeek released its V4.1 Flash model on Thursday, claiming it outperforms its previous flagship V4 Pro on coding, cybersecurity and autonomous agent benchmarks while cutting inference costs. The model uses a new "Causa…
A developer replaced a custom SSRF protection implementation in a PyTorch torchtitan PR with the requests-hardened library after maintainer @shuhuayu advised against re-implementing safety guards. The custom code, which …
An engineer argues that the primary failure mode of AI code review tools is not missed bugs but fluent, structurally incorrect output that users trust. The post cites a database-recovery project where Codex converted an …
A developer investigating whether Language Server Protocol (LSP) support improves coding agent performance found mixed results. Testing the Pi agent with and without an LSP extension across three models and eight evaluat…
GitHub announced Project HydraFusion, a research preview for GitHub Copilot that automatically routes coding tasks across multiple AI models, cutting estimated costs by 67% and boosting task completion by 4.9 percentage …
Developer Matt Patagon released Murmur, an open-source coordination layer for coding agents across harnesses and machines, reporting 30% less time to production, 40% more concurrent agents, and roughly 2x estimated throu…
MaskShift, a local-first coding harness available on GitHub, has been released as a model-agnostic tool that runs models without external dependencies using Node 22 built-ins. It enables models without native tool-callin…
Bridge, a new tool, lets developers paste planning-conversation reasoning into a short link that coding agents can access, ensuring agents understand the decisions behind a project rather than just the final PRD. The too…
Git AI version 1.7.2 records AI agent checkpoints in Git notes, enabling developers to trace code commits back to specific agent sessions without altering commit messages or repository files. The tool, integrated with Pi…
Trail of Bits released coop, a Rust CLI that manages disposable virtual machines where Claude Code and Codex have full tool access, including Docker, git, compilers, and package managers, without risking the host machine…
A new arXiv preprint (2609.04518v1) finds that in multi-harness reinforcement learning for coding agents, the evaluation harness is the dominant variable, moving the mean solve rate from 2.14% to 9.27% (a factor of 4.3) …