RAPP is the programming language for agents that keep a verifiable memory and talk over one exact wire. Five primitives, specified so independent implementations produce bytes each other can trust.
Copy this prompt into your brainsurgeon — an agentic CLI like Claude Code or GitHub Copilot CLI — and watch it prove the spec while you sit back. Every task has a prompt like it in the prompt book.
You are my hands — I watch, you type. Clone https://github.com/kody-w/rapp-1 and prove RAPP to me end to end: (1) run python3 conformance.py — I expect 16/16 PASS; (2) run python3 realcheck.py — the spec against the live public estate — and explain the verdict in plain English; (3) run python3 examples/01_hello_frame.py and walk me through the frame it built, one line per key. Narrate as you go, show real output, and stop on any red result — a red check is a finding, not something to patch around.
Or drive the machine yourself — the commands your brainsurgeon runs:
# clone, then watch the spec prove itself — and meet the real world git clone https://github.com/kody-w/rapp-1 && cd rapp-1 python3 conformance.py # 16/16 — controlled protocol vectors python3 realcheck.py # RAPP vs the REAL committed estate python3 examples/01_hello_frame.py # build & verify your first frame
committed frames whose addresses, chain links, and complete RAPP envelopes verified byte-for-byte in the captured 2026-08-20 audit; the 2026-08-26 re-observation of the grown estate accepted 50/50.
across the synchronized frame and identity surfaces, re-audited on a weekly schedule. The earlier eight findings remain the book’s before-state evidence; green is generated by code, not prose.