thinking
A Claude Code plugin marketplace of thinking and reasoning skills, built on the open Agent Skills standard - so the same skills work in Claude Code, GitHub Copilot CLI, and other compatible agents.
Skills
- consciousness-council - multi-perspective "Mind Council" deliberation on decisions and creative challenges.
- scientific-brainstorming - creative research ideation, interdisciplinary connections, and gap-finding.
- context-degradation - recognize and mitigate context failures (poisoning, clash, lost-in-middle).
- context-fundamentals - foundational context engineering for AI agent systems.
Install in Claude Code (plugin)
/plugin marketplace add corticalstack/thinking
/plugin install thinking@thinking
Restart Claude Code. The skills then appear in autocomplete as /thinking:<name> (e.g. /thinking:consciousness-council).
Install in GitHub Copilot CLI (and other agents)
The skills live in .claude/skills/, which Copilot CLI reads. Install them with the gh skill command (GitHub CLI v2.90.0+):
# all skills from this repo, available everywhere (user scope):
gh skill install corticalstack/thinking --agent github-copilot --scope user
# or just one skill:
gh skill install corticalstack/thinking consciousness-council --agent github-copilot --scope user
In a Copilot CLI session: reload with /skills reload, confirm with /skills info consciousness-council, then invoke by name, e.g. "Use the /scientific-brainstorming skill on this research question."
gh skill also accepts --agent claude-code, cursor, gemini-cli, codex, and many others, so the same repo serves any Agent-Skills-compatible tool. (Alternatively, copy a skill's folder into your project .claude/skills/ or personal ~/.copilot/skills/ and run /skills reload.)
Licensing
Repository scaffolding (manifests, README) is MIT. Individual skills retain their own licenses where present: consciousness-council and scientific-brainstorming are MIT; context-degradation and context-fundamentals carry no explicit license - treat them as the maintainer's own.