Inspired by caveman, I built a skill to do the same things with more tokens

I made a prompt that fixes AI-written content.

I built a structured context layer for AI coding agents so they stop generating the wrong UI

CLI tool to edit google docs using markdown - granular edits via diff / apply based workflow

Tool for inline annotation of LLM-generated specs and prompts (works with any MCP client)

Stop pretending you're in control. You're just a high-end biological battery

Arc Gate — LLM proxy that catches 100% of indirect/roleplay prompt injection attacks (beats OpenAI Moderation and LlamaGuard)