J-Space provides a selective workspace designed for tasks requiring durable state and cross-file consistency. It enables AI agents to maintain evidence and recovery paths during complex reasoning, coordinated agent workflows, and authorized security analysis.
This suite helps automation builders implement inference-time control for long-horizon work, ensuring that repository engineering remains structured and verifiable without overloading direct responses.
Follow the documented setup, then try a first task.
Installs: Complete j-space suite, including SKILL.md, modules, references, and scripts.
Run from: The task directory, unless a controller is invoked with --root TASK_DIRECTORY.
Copy the complete j-space/ directory into the Skills directory specified by the host. Use an empty destination, preserve SKILL.md, modules/, references/, and scripts/ together, and avoid creating an extra j-space/j-space/ nesting level. When distributing the standalone skill, also place LICENSE and THIRD_PARTY_NOTICES.md beside the installed SKILL.md.
In a terminal, verify the installed directory. Replace <python-command> with an available python, python3, or py -3 command, and replace <skill-root> with the installed j-space directory.
<python-command> <skill-root>/scripts/verify_suite.py
Reload the host if it discovers skills only at startup, select j-space through its skill UI, and confirm that it can retrieve a routed module. Use $j-space or /j-space only when the host documents that syntax; otherwise ask the host to read the installed SKILL.md explicitly.
Uses j-space
Use j-space to modify this repository. Inspect the existing contracts, maintain a source-backed map, delegate independent work where useful, and verify the final behavior.
Claude