J-Space Cognition Suite

J-Space Cognition Suite is a model-agnostic, inference-time control system for deep reasoning, long-horizon work, tool use, verification and recovery, packaged as a Skill for cross-platform, selective installation into an AI agent host. It operates entirely at inference time, leaving model weights and training unchanged, and organizes an agent's working representations into a deliberately managed workspace through one entry point, nine selectively loaded modules, four reference documents and an optional controller for durable task state.

Installation copies the suite's directory into a host's Skills folder and runs a bundled Python integrity check before the host reloads and discovers it. An entry gate automatically selects the lightest of three operating passes for a given request: a fast pass for a one-step or easily checked task, a full pass that loads one or two relevant modules for several dependent steps, and a loop pass that adds a ledger, checkpoints and recovery for long, multi-file, multi-turn work.