Browse
Latest
Newest published posts first. Popular ranks by views.
Evaluate AI Coding Agents at the Patch, Not the Promise
A practical system for evaluating AI coding-agent changes with task contracts, fail-to-pass tests, pass-to-pass regressions, and diff review.
Four boundaries to set before handing a repository to an AI agent
A practical guide to scope, context, authority, and verification when using AI agents on a code repository.
Don't Kill Auto From the Email Alone — Cursor's Aug 24 Per-Model Pricing
From 24 Aug 2026, Cursor Auto bills at the routed model's list price instead of a flat Auto rate. Included limits rose; routing did not change — check Spending before you kill Auto.
Don't Leave JetBrains Outside the Policy — Copilot Enterprise Managed Settings
GitHub Copilot for JetBrains now honors enterprise managed settings (2026-08-18). Close the IntelliJ gap for MCP allowlists, plugins, OpenTelemetry, and Bypass/Autopilot.
Don't Run That headersHelper Blind — Claude Code 2.1.238 Moves the MCP Trust Boundary
Starting in Claude Code 2.1.238, project and plugin headersHelper commands require folder trust and strip credential-like env vars. Treat shared .mcp.json as a shell execution surface — with a practical checklist.
Before You @GitHub in Standup Chat — Teams Shared Agents and the Trust Boundary
Teams Copilot shared sessions (2026-08-21 preview): write-access triggers, sandbox + AI credit budgets, extra PR approval for the Teams identity — companion to Slack Code, not a rehash.
Before @GitHub Ships From Slack — Code Channels and the Trust Boundary
GitHub’s 2026-08-21 Slack agentic preview. Lock write access, app-identity PRs, extra approvals, and cloud-agent budgets before the first @mention.
When the Agent Wakes Itself — Treat Cursor Subscriptions and /goal as a Harness
Cursor’s 2026-08-19 cloud agent release: Subscriptions, /goal, and isolated subagent VMs. How to run always-on loops with stop conditions—not autopilot.
Don't Code the First Mock — What Claude Code /design Actually Changes
Claude Code's /design research preview (≈2026-08-17) adds editable UI artboards before implementation. Treat it as a review gate — not 'design system solved.'
Cursor Is AIUC-1 Certified — What That Evidence Buys Security Teams (and What It Doesn't)
Cursor’s 13 Aug 2026 AIUC-1 certification is recurring adversarial evidence for coding-agent risk—not a license to drop Run Mode, MCP allowlists, or .cursorignore. Covers Schellman, A008/B010, and operator baselines.
Don't Rip Out GitHub on Day One — What Cursor Origin's Early Beta Actually Ships
Aug 17, 2026 Origin early beta ships repos, PRs, browse, and GitHub sync. Agent-native hosting is soon. Keep GitHub as SoT while synced; Issues/Actions stay on GitHub.
Don't Install That Portable Plugin Blind — Agent Plugins 1.0 Moves the Trust Boundary
Agent Plugins 1.0 standardizes packaging for skills and MCP — not install trust. Separate Cursor Plugins from portable plugins, and lock Copilot managed-settings before opening marketplaces.
Cursor Is Part of SpaceX Now — What to Change in Your Agent Stack (and What Not To)
2026-08-14 is an org close. Update pickers, privacy checks, and Copilot admin gates from shipped product notices — not from the $60B headline alone.
Don’t Flip On MAI-Code-1.1-Flash From the Changelog Alone — Watch the September 10 Clock
MAI-Code-1.1-Flash rolled out in Copilot on Aug 11. Business policy is off by default; MAI-Code-1-Flash retires Sep 10. Don’t mix 0.25× with usage-based list prices.
Cursor Cloud Agent Builds — Warm the Environment Before the Agent Thinks
Builds (Aug 13) pre-bakes Cloud Agent environments; default on Aug 17. Split install vs start, keep last-good builds, and keep user secrets out of snapshots.
Claude Code Flips to Auto Mode on August 14 — Set Your Boundaries Before It Does
Pro, Max, and Team new sessions default to auto mode. It is a classifier gate—not bypass. Pin ask rules and autoMode.environment before the flip.
Minimize the Chat Without Losing the Thread — Copilot Web Conversation Controls
Aug 10 Copilot web chat adds minimize/resume and token spend indicators. Mid-session credit hygiene meets the Aug 11 per-model usage report.
Code Quality No Longer Smuggles Copilot Onto Your PRs
As of 2026-08-07, enabling Code Quality no longer auto-creates a Copilot review ruleset. Audit leftover toggles, opt in explicitly, and set Lite vs Balanced depth on purpose.
When a Comment Wakes the Agent — Trust Handoff After Black Hat 2026
In Black Hat USA 2026 week, comment-triggered Copilot automations landed beside research showing GitHub text can hijack coding agents. Treat comments as untrusted tool input.
Kimi K3 Is GA in Copilot — Don’t Flip It On From the Changelog Alone
GitHub made Moonshot’s Kimi K3 generally available in Copilot on 2026-08-06 — then paused and resumed the rollout the same day over a GitHub Actions incident. Business/Enterprise stays off by default. Verify picker, policy, pricing, and open-weight risk notes before a bounded trial.