Instruction-layer protocols that can be inspected, versioned, forked, tested, and reused across AI assistants. The Portable Reasoning and Research Intelligence protocols ship as SKILL.md-based Skills, with installation guides for ChatGPT, Claude, Gemini, and GitHub Copilot.
01
Skill · v1.0
Portable Reasoning Protocol
A portable reasoning layer for more rigorous, evidence-bounded AI work.
Adapts its rigor to the task: simple work stays simple; consequential, ambiguous, or difficult-to-reverse work receives deeper scrutiny. It eliminates invalid transitions between ideas, evidence, claims, and actions.
Adaptive rigorFour invariantsHallucination control
Who it is for
Founders, executives, analysts, researchers, product and engineering teams, legal, risk, and governance professionals — anyone using AI for consequential decisions.
02
Skill · v1.0
Research Intelligence Protocol
Disciplined discovery and cross-domain structural abstraction.
Two components kept deliberately distinct: the Discovery Protocol turns reality into structured evidence; the Abstractor of Abstractors turns evidence into tested, transferable invariants.
Discovery ProtocolAbstractor of AbstractorsEquivalence ladder
Who it is for
Researchers, R&D and engineering teams, architects, consultants, due-diligence, technology-transfer, patent, and root-cause analysis teams.
03
Skill · S43
TFA Protocol
Say what is true. Ask for nothing. Protect their next move.
A lightweight, optional behavioral protocol for AI interaction, built on three rules — Truth, Freedom and Agency — with implementation and evaluation guidance that explains the rules without turning them into an authority system.
TruthFreedomAgency
Who it is for
Teams building or reviewing AI assistants who want a small, shared vocabulary for judging interaction quality — calibrated claims, no unnecessary extraction, and user choice kept open.