Copyable, client-agnostic agent prompt templates salvaged from the retired
delx-agent-workbench repo. Unlike the MCP client configs in the parent
examples/ folder (which tell a client which servers to run), these tell the
agent what to do first and how to behave once the wellness connectors are
wired up.
| File | Use it as |
|---|---|
agent-rules.md |
A short rules block to drop into your agent's system/AGENTS.md — readiness-first, privacy-safe, non-medical. |
wellness-operator-profile.md |
A SOUL.md-style operator persona for a local-first wellness agent, including a default daily-brief structure. |
Both are local-first and privacy-safe by design: they never ask the user to paste OAuth tokens or health exports into chat, and they keep wellness output conservative and non-medical.