Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/xxyoudeadpunkxx/sensecraftxstudio/llms.txt

Use this file to discover all available pages before exploring further.

SensecraftXStudio is a single reusable file — AGENTS.md — that acts as an active operational contract for AI coding agents. Rather than offering passive documentation or a loose collection of prompts, it defines how every response must be framed, bounded, and reported. The contract is built in layers so that each section addresses a distinct phase of an agent’s work: from first activation, through pre-action framing and execution discipline, to final reporting.
The contract has layered force — some sections govern pre-action framing before any move is chosen, others govern execution and reporting after the move is closed. Some rules do not specify an action at all: they govern the baseline reading from which action is later derived.

Layer Guide

Each layer of the contract handles a distinct concern. The table below maps layer names to their roles in a session.
LayerRole
Contract Activation / Scope / DomainWhat this contract governs and when it is in force
Operating Posture / Horizontal Plane / early invariantsHow task framing must begin, and how the task must be read and bounded before action
Later invariants / Stop ConditionsHow action is bounded, checked, or refused
Output / Final Response ContractHow results must be produced and reported

Contract Sections

Activation

How to activate the contract, what drift looks like, and the re-entry procedure when drift is detected.

Operating Posture

The pre-action posture the contract requires — read the system before acting on the task.

Horizontal Plane

The three A/B/C axes that every move must pass through before and during execution.

Stop Conditions

The conditions under which the assistant must halt and surface the conflict before continuing.

Final Response

The four required fields — Touch, Ground, State, and Convergence — that close every consequential response.

Build docs developers (and LLMs) love