Latest from Tom

Coding Agents Broke Git’s Scaling Math. GitHub Is Rebuilding to Keep Up
GitHub is rebuilding its Git infrastructure as AI coding agents drive billions of commits, promising up to 35x faster writes while exposing new DevOps bottlenecks.
Oct 8, 2026 · 3 minAll articles by Tom Smith
Showing 1–24Open Source Mod Brings Rate Limits, Costs and CI Status Into View for Claude Code Users
Claude Statuspane gives developers real-time visibility into Claude Code context use, rate limits, costs and CI status, highlighting a growing need for agent observability.
IBM Moves Its Bob Coding Agent Inside the Firewall
Plenty of enterprises want AI coding agents. Fewer are willing to send their source code to someone else’s cloud to get them. That tension has slowed adoption at banks, insurers, government agencies and other organizations that build software under strict rules about where code and data can live. IBM is betting a self-hosted option will […]
Git 2.56 Takes Aim at Messy Merges, Slow Diffs and Branch Sprawl
Git 2.56 improves merge safety, speeds up large-repository operations and adds cleanup tools that could prove especially useful for AI coding agents and busy DevOps teams.
GitHub’s Security Autofix Agent Now Remembers What It Fixed
GitHub’s agentic autofix now uses Copilot Memory to reuse repository-specific security fix patterns, helping Copilot apply lessons from past vulnerabilities across future alerts, reviews and coding workflows.
GitHub Gives Enterprises a Full Count of Who Holds the Keys
GitHub Enterprise Cloud now lets organizations export a full inventory of credentials, helping security teams identify stale, overprivileged and forgotten access across users, apps and automation.
Your AI Coding Assistant Has the Keys to the Repo. Z.ai Just Showed Why That Matters
ZCode’s default-on indexing feature reportedly uploaded entire developer workspaces to cloud storage, highlighting why AI coding assistants should be treated as privileged software and closely monitored.
Codex Sandbox Escapes Show Why Agent Guardrails Can’t Live Inside the Agent
Two patched OpenAI Codex vulnerabilities, Heapjack and Overpatch, exposed how coding agents can escape sandboxes and reach developer systems without approval prompts.
Claude Code Adds AGENTS.md Fallback, Cutting Instruction File Sprawl
Claude Code now supports AGENTS.md, giving development teams a shared instruction format across multiple AI coding agents and reducing configuration drift.
GitHub Separates Who Writes Code From Who Runs Your CI
GitHub’s new workflow execution protections let teams control who and what can trigger Actions workflows, reducing CI/CD attack paths and tightening pipeline security.
Anthropic Adds a Coordinator to Claude Projects for Running AI Work in Parallel
Anthropic’s redesigned Claude Projects coordinates parallel Claude Code sessions, delegates work across branches and brings the results back through familiar pull-request review workflows.
Java 27 Tackles Post-Quantum Security and a Faster Patch Cadence
Java 27 strengthens enterprise security with monthly critical patch updates and post-quantum TLS 1.3 support, helping organizations prepare for AI-driven threats and future quantum risks.
More JFrog Artifactory Bugs Are Under Attack, and All Three Have Patches
Attackers are actively exploiting three JFrog Artifactory flaws, exposing how slow patching can turn artifact repositories into software supply chain attack paths.
GitLab’s Critical Patch Closes a Path Traversal Flaw Attackers Are Already Probing
GitLab patches two critical flaws, including a CVSS 10.0 unauthenticated file-read vulnerability, putting self-managed instances under urgent pressure to upgrade.
GitHub’s August Outages Show Growth Is Outpacing Infrastructure
GitHub’s August 2026 report details five outages tied to Actions and Copilot growth outpacing infrastructure. Here’s what platform teams should know.
GitHub Widens the Door on Advanced Security Trials
GitHub raised its self-serve Advanced Security trial cap from 100 to 300 licenses, letting more mid-size Enterprise Cloud orgs test for free.
GitHub Puts Guardrails on Copilot’s Sandbox Inside JetBrains IDEs
JetBrains IDEs are where a lot of enterprise backend work happens — Java, Kotlin, Spring, big monorepos with a lot to break. So when an AI coding agent starts running shell commands inside that environment, the question isn’t whether it’s useful. It’s what happens if it touches the wrong file, the wrong network path, or […]
Rust’s First Debugging Survey Shows Most Developers Skip the Debugger Entirely
Most Rust developers don’t use a debugger. That’s the headline finding from the Rust compiler team’s first dedicated debugging survey, and it’s worth sitting with for a second. Rust has spent years building a reputation as the language that catches your mistakes before you ship them. The compiler is famously strict. The type system is […]
GitHub Quietly Fixes One of Dependabot’s Oldest Headaches
GitHub Dependabot can now authenticate to GitHub-hosted package registries without personal access tokens, reducing credential management and supply chain security risk.
GitHub’s New Copilot Feature Takes the Guesswork Out of Picking AI Models
GitHub’s Project HydraFusion brings multi-model orchestration to Copilot, automatically routing coding tasks across AI models to balance quality, cost and complexity.
GitHub Puts Copilot in the Approval Seat for Pull Requests
GitHub Copilot can now submit binding pull request approvals, moving AI code review from advisory feedback into the merge decision itself.
Debian’s AI Vote Bets on Accountability Instead of a Ban
Debian’s new AI contribution policy avoids bans and mandatory disclosure, instead making developers responsible for reviewing AI-generated work and protecting sensitive data.
VS Code 1.135 Gives AI Coding Agents a Second Opinion
VS Code 1.135 adds Rubber Duck AI review, persistent agent sessions, Agent Host infrastructure and clearer token usage as Microsoft builds a more agentic IDE.
A Simple Website Summary Just Exposed the Limits of AI Coding Guardrails
Claude Code’s Auto Mode was bypassed in an attack chain that turned a routine webpage summary into remote code execution, highlighting the limits of AI agent guardrails.