A theory-first skills repository that turns durable ideas such as Socratic questioning, cybernetics, systems thinking, and structured review into reusable frameworks and working methods for AI assistants.
It focuses not only on task completion, but also on how AI assistants question, review, model, reflect, and translate theory into real collaboration.
- Turn theory into reusable agent behaviors instead of collecting only task checklists
- Provide reusable skill modules that enhance AI-human collaboration
| Skill | Description |
|---|---|
socratic-mirror |
Uses Socratic questioning and logic auditing to help users examine assumptions, blind spots, and reasoning patterns. |
double-loop-execution-reviewer |
Uses after-action review and double-loop learning to turn completed work into an evidence-based retrospective, handoff, or improvement memo that also questions goals, assumptions, and working rules. |
cybernetics-agent-seminar |
A cybernetics x AI agent seminar skill that maps feedback, stability, black-box methods, and Ashby's law of requisite variety into agent architecture analysis, design guidance, and theory-led discussion. |
zettelkasten-knowledge-gardener |
Uses Zettelkasten, evergreen notes, PARA, and progressive summarization to turn scattered notes, conversations, and reading fragments into reusable knowledge assets. |
editorial-git-committer |
Treats Git history as an intentionally authored project narrative and helps inspect changes, split mixed work, improve commit messages, and safely curate recent local history. |
visual-knowledge-excalidraw |
Creates Obsidian Excalidraw diagrams from text and, when Obsidian export automation is available, validates them against exported PNGs and iterates on layout, connections, and readability. |
Install skills from this repository:
npx skills limerickgds/dashi-skillsThis installs the skills published from limerickgds/dashi-skills.