Tools I give away.

Cuts context bloat, routes AI to the right model, saves money on API costs, and helps Claude work more autonomously without hand-holding.

Stop installing packages blind. GuardDog scans npm and PyPI packages for malware, live threats, CVEs, and suspicious patterns before you install. Free, one command to set up.

One routing layer, one capability map, and thin host adapters so the same skill library works across Claude Code, Codex, Gemini CLI, and adjacent runtimes.

A Claude Code skill that runs a believability-weighted decision council. Five scouts argue your question, a synthesizer weighs them, and you get PROCEED, HOLD, or INVESTIGATE with a confidence score.

Strip AI writing patterns and rewrite in your actual voice. Learns from your podcasts, books, and content.

A practical checklist for keeping Claude Code and OpenClaw setups low-risk and compliant, plus an audit prompt to scan your own stack for policy-triggering patterns.