BlogWikiAbout

Kyle Pericak

"It works in my environment"

AI-Native SDLC: First Try

AI-Native SDLC: First Try

Created: 2026-03-16

Building a product development lifecycle where AI agents handle scoping, research, and documentation, from PRD through design doc to implementation plan.

My First Cron/Event-Triggered AI Agents on K8s

My First Cron/Event-Triggered AI Agents on K8s

Created: 2026-03-15, Modified: 2026-03-15

Building a K8s controller that runs a Claude Code agent every morning to search the web for AI news and write a digest to the wiki. Also triggerable on demand via webhook.

Applying Research to Claude Agents

Applying Research to Claude Agents

Created: 2026-03-15, Modified: 2026-03-15

Research-backed guidance on getting coding agents to do useful work, applied to rebuild a 7-agent roster from Anthropic's published advice and three deep research reports.

Cloudflare MCP: Query DNS and the Full API

Cloudflare MCP: Query DNS and the Full API

Created: 2026-03-12, Modified: 2026-03-15

Connecting Cloudflare's remote MCP server to Claude Code so it can query DNS records, zone configs, and the full Cloudflare API from the terminal.

Attempting an AI Agent Org Chart

Attempting an AI Agent Org Chart

Created: 2026-03-11, Modified: 2026-03-12

Experimenting with multi-agent setup with named roles, a shared wiki, and an orchestration agent that coordinates between them.

Safety Hooks for Claude Code

Safety Hooks for Claude Code

Created: 2026-03-11, Modified: 2026-03-11

Three shell scripts that block destructive commands, protect sensitive files, and log everything Claude does.

OpenRouter for Claude Code, OpenCode, and OpenClaw

OpenRouter for Claude Code, OpenCode, and OpenClaw

Created: 2026-03-10, Modified: 2026-03-10

Configuring OpenRouter to be where all my tools go for tokens and making a custom MCP server for it

Lint Toolkit for AI Agents

Lint Toolkit for AI Agents

Created: 2026-03-07, Modified: 2026-03-07

A Docker image bundling ruff, biome, hadolint, and tflint so AI coding agents can lint Python, JS/TS, Dockerfiles, and Terraform.

Security Toolkit for AI Agents

Security Toolkit for AI Agents

Created: 2026-03-07, Modified: 2026-03-07

A Docker image bundling semgrep, trivy, and gitleaks so AI coding agents can run security scans as part of their workflow.

Linear MCP: Planning with Robots

Linear MCP: Planning with Robots

Created: 2026-03-06, Modified: 2026-03-06

Connecting AI coding tools directly to Linear so AI can pull tasks in

GA4 and Its MCP Server

GA4 and Its MCP Server

Created: 2026-03-04, Modified: 2026-03-05

Adding Google Analytics 4 to a statically-exported Next.js blog, then connecting the official GA4 MCP server so Claude can query the data directly.

Playwright MCP: Claude Code Verifying Its Own Work

Playwright MCP: Claude Code Verifying Its Own Work

Created: 2026-03-03, Modified: 2026-03-03

Using the Playwright MCP server to give Claude Code live browser access, writing Gherkin-driven tests for a Next.js blog, and exploring what it means for an agent to verify the changes it just made.

Cross-Tool AI Rules with Ruler

Cross-Tool AI Rules with Ruler

Created: 2026-03-02, Modified: 2026-03-02

Using Ruler to maintain a single source of truth for AI coding assistant instructions across Cursor, Claude Code, and OpenCode in a monorepo.

Claude Code + CodeRabbit: Review Before the PR

Claude Code + CodeRabbit: Review Before the PR

Created: 2026-02-28, Modified: 2026-03-11

Setting up the CodeRabbit CLI and Claude Code plugin so you can run AI code review locally before creating a pull request, and configuring both tools to agree on your project's standards via CLAUDE.md

Blog code last updated on 2026-03-18: dc88d520b39a657ab6232142582c365c23941b60