J-Space Cognition Suite skill
An inference-time reasoning control skill: deep analysis, long tasks, verification and recovery grounded in durable evidence
- 2,956 ★ on the repository
- Last commit
- Apache-2.0
Extended access
Why this level
- Runs Python controllers and persists task state on disk
- Reads the project's source text for checks
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/j-space-cognition-suiteDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Runs Python controllers and persists task state on disk
- Reads the project's source text for checks
This is third-party code. Review the repository files before installing.
What it does
J-Space is a model-agnostic skill for controlling an agent's reasoning during a task. It installs as one SKILL.md with thirteen modules and a single routing entry, and standard-library Python scripts persist state, reread the source text, detect stale maps and evidence and return a blocking result when a required condition is missing. There are four operating levels, from a fast check to a loop with bounded agents, a second consideration and independent review. The task is stated together with acceptance conditions, and the agent keeps long-task decisions connected to verifiable evidence. The idea draws on global workspace research.
Who it is for. For engineers and teams who run complex, multi-stage work through an agent with dependencies, verification and recovery.
Good fit when
- You have long, multi-file work needing verification at each step
- You want the agent's decisions tied to evidence
- You need agent coordination and independent review for complex integration
Not a fit when
- The task is simple and checkable at a glance
- The host cannot load a local SKILL.md and its files
Example request
Use j-space at high level to rework this repository with a source-backed map and verification of the final behaviorLimitations
It needs a host that can load a local SKILL.md and its files; executable controllers and validation need Python 3.10 or newer, while low and medium levels work without it via a text fallback. Installing does not register hooks, launch agents or grant tool access by itself; the host does that. The skill guides the process, but the outcome depends on your task and agent.
How to disable. Remove the j-space folder from the host's skills directory.
FAQ
Do I need pip?
No. The scripts use the Python standard library with no dependencies or background service; a 3.10+ interpreter is needed for the controllers.
What do the levels give?
From a fast check at low level to a loop with shared state, bounded agents and independent review at high level.
Discussion
No questions or reviews yet.
Related
SkillsA skill that convenes a council of thinker personas and runs a structured deliberation on a hard decision, ending with a verdict
A parallel divergent ideation skill for agents: fans out under cognitive frames, prunes traps and deepens the best branches
A completion-discipline skill for large agent tasks: write the acceptance ledger with runnable checks first, then do the work, then reverify against evidence
A skill plus hooks that keep a coding agent on task and cut speculative defenses, unread checksums and scope creep
A skill that turns several Claude Code, Codex and Cursor windows into a team with a coordinator, shared registry and cross-review
A skill with 25 commands to audit and improve the prompts, tools and agent designs of a project in Cursor, Claude Code and Codex
README in short
The README describes J-Space as an inference-time reasoning control skill for complex analysis, repository engineering, agent coordination and authorized security analysis. One skill with thirteen modules and a shared routing entry, with Python scripts that persist state, reread source and block when required conditions are missing. Install by copying the j-space folder into the skills directory, verify with verify_suite.py, and use four operating levels from a fast check to a team loop. The idea draws on global workspace research, Apache-2.0 licensed.