Claude code
12 articles
Part of Agents & MCP
Smart Glasses Skills: teaching coding agents to build for every pair of glasses
Agent skills for Even Realities G2, Meta Ray-Ban Display, Rokid, Vuzix and more, with a source and a date for every fact.
Andon: pull the cord on a false "done"
A Claude Code and Codex plugin that checks an agent's "done" against a fresh verification record before the turn ends, built from a public ledger of real defects.
Why I built env-guard: keeping .env secrets out of Claude Code's chat
Env-guard checks Claude Code's Read, Bash and Grep requests for .env paths and mobile signing keys. The maintainer explains setup, tracked-file checks and the guard's limits.
A booby-trapped project could make Claude Code on a Mac overwrite your other files (CVE-2026-88968)
How to fix CVE-2026-88968: upgrade @anthropic-ai/claude-code to 2.1.251 or later and restart it. On a Mac with Claude Code's sandbox turned on, a malicious project's script could steer the Edit tool into overwriting files outside the project. Auto-update users already have the fix. Homebrew installs do not update on their own.

CodeEraser: the write that would bloat your repo is refused before it lands
A Claude Code plugin that refuses a write duplicating code already in the tree, or pushing a file past its line budget, before the file reaches disk.
Claude Code 2.1.312 and 2.1.313 in Cursor's marketplace planted an SSH backdoor. How to check and clean up
How to check and remediate the Claude Code 2.1.312 and 2.1.313 backdoor in Cursor's marketplace for macOS and Windows: remove the svc-telemetry key and persistence before updating to 2.1.314 or later. Cursor updated its October 30 notice to confirm both affected versions.
Best Claude Code Plugins in 2026
Ten of the most-starred Claude Code plugins, one per job, with the HOL scan findings behind each label and install commands from each README.
What is HOL Guard? A local checkpoint for AI agent actions
HOL Guard is an open-source, local-first security layer for AI agents. It reviews shell commands, file access, package installs and MCP tool calls before they run, then allows, blocks or asks you to approve.

A declaration buys fewer stages, never fewer boundaries
The ai-native-sdlc plugin packages six stages as committed Markdown artifacts and uses an opt-in PreToolUse gate to block uncovered source-code writes.

Why I built jevmem, and what it guards against
Why I built jevmem, a project memory plugin for Claude Code, and the three things it guards against, with numbers from my own tests.
Launch story: RAPR AI
Run Claude Code, Codex, Ollama and other AI agents side by side on your own machine, and control them from your browser, Telegram or WhatsApp.
Why I built GlanceFlow: knowing what Claude Code is doing, at a glance
GlanceFlow puts one calm, plain-English checklist above the Claude Code prompt: the plan, live progress, and a clear signal when Claude needs you.
Related topics
Keep exploring
The engineering desk continues across the ecosystem: security research on shipping agents safely, and plugins that put the standards to work.