Personal AI system prompt / custom instructions block, maintained for use across AI platforms (Claude, ChatGPT, Perplexity, etc.).
This repo tracks the evolution of my personal AI custom instructions over time. The instructions define how AI assistants should behave, format responses, and interact with me across all general-use AI platforms.
These are not project-specific or task-specific prompts. They establish baseline behavior, tone, formatting standards, and workflow conventions that apply to every interaction.
Paste the contents of the current instructions file into the Custom Instructions, System Prompt, or Personal Preferences field of any AI platform. Intended for general-purpose, non-isolated chat contexts.
Platforms this is used on:
- Claude (claude.ai)
- ChatGPT
- Perplexity
Each version is saved as a new dated file. The most recent file is the active version.
- Tone & Behavior — How the AI should communicate (direct, professional, no filler)
- Formatting — When to use prose vs. headers vs. tables vs. code blocks
- Clarifying Questions — When to ask before answering vs. after
- Coding Tasks — Planning-first workflow, approval before generating code
- Proactive Flags — When and how to surface related considerations I might miss
- Writing & Creative Tasks — Audience clarification before proceeding
- Environment & Tooling — Platform and productivity stack context
- Context File — Instructions for the AI to prompt updates as preferences evolve
New versions are created when meaningful changes are made — refined preferences, new instructions, or behavior corrections based on real-world use. Commit messages should briefly note what changed and why.
Developed iteratively through a structured conversation designed to capture working style, communication preferences, technical background, and workflow conventions. The goal is an AI that behaves like a knowledgeable professional colleague — accurate, direct, and efficient.