[FEATURE] Native support for user governance files (user-defined AI terms)

What

Support for a user governance file that Claude loads automatically at every session start. The user defines their rights, rules, session protocol, and learning loop in a single file. Claude reads it before responding.

Why

Currently, users who want to define how AI treats them must paste a markdown file into every session manually. This is high-friction and breaks continuity.

CLAUDE.md already exists for project-level instructions. This request is for the equivalent on the user side: a persistent file where the user defines their terms for the relationship, not just the project.

What it would look like

• A file (e.g., MY-TERMS.md or USER-TERMS.md) in a standard location
• Claude reads it at session start, alongside CLAUDE.md
• The user's terms and Anthropic's terms coexist. Neither overrides the other. Conflicts are named transparently
• The file is portable: works across Claude Code, Claude chat, and ideally any AI

Context

I built a framework for this: On Your Terms AI (GitHub). One file. Rights, rules, a learning loop. Tested daily on Claude Code for months.

It works. But it depends on the user bringing the file. If Claude had native support, user-defined terms become infrastructure instead of friction.

The ask

Consider a native mechanism for users to define persistent terms that Claude respects across sessions. Not settings. Not preferences. A governance file the user writes and owns.