Small enough to begin
Start with Raw, Wiki, and Outputs. Add modules only when your actual work needs them.
Open source · Codex · Markdown · Obsidian
A knowledge system that grows with you
Eleven Codex skills for students, researchers, and knowledge workers. Capture scattered information, connect atomic notes, preserve project evidence, review experience, and observe personal patterns without rushing to fixed conclusions.
More than a place to save notes
The system is designed around what happens after capture: understanding, evidence, action, reflection, and reuse.
Connected with intent
Atomic notes link to meaningful sources, topics, dependencies, and applications rather than decorative backlinks.
Personalized
An onboarding interview considers goals, habits, existing files, and privacy boundaries before proposing a structure.
Built for growth
Resources are inputs. Capability is supported by explanation, practice, feedback, exception handling, and transfer evidence.
Ready for real projects
Follow questions, evidence, decisions, execution, validation, retrospectives, delivery, and later updates.
Careful self-observation
Keep context, confidence, counterevidence, and alternative explanations. One event never becomes a personality verdict.
A reproducible knowledge workflow
Use one stage or follow a learning experience or project through the complete loop.
CapturePreserveExamineConnectPlanPracticeValidateReviewDistillPublishUpdate
Eleven independently callable skills
Call $knowledge-vault when you do not know where to begin. It routes your request to the shortest relevant workflow.
$knowledge-vault
Routes requests across setup, intake, organization, growth, projects, output, and review.
$vault-bootstrap
Interviews needs and privacy boundaries before creating a minimum viable vault.
$knowledge-intake
Processes mixed material while preserving sources and distinguishing fact, judgment, and inference.
$atomic-notes
Creates focused reusable notes and searches for existing knowledge before adding another file.
$obsidian-vault
Adds Obsidian workflows while keeping the underlying Markdown portable.
$growth-system
Assesses learning candidates, capacity, practice, and capability evidence.
$inspire-time
Connects a small transferable insight to an observable real-world practice.
$project-lifecycle
Tracks goals, decisions, evidence, interruption recovery, validation, and completion.
$retrospective
Reviews projects or periods from recorded events and turns learning into a next experiment.
$self-observation
Examines recurring preferences and work patterns without diagnosis or fixed labels.
$knowledge-output
Creates audience-specific work with source versioning and disclosure control.
Work, research, learning, and life
The same evidence-aware structure can support a long project or hold an idea that is not ready for commitment.
Projects
Preserve scope, decisions, execution evidence, interruption state, acceptance, and review.
Research
Connect questions, sources, hypotheses, experiments, failures, and provisional findings.
Learning
Move from resources toward explanation, practice, feedback, and transfer.
Personal development
Keep small moments of insight, candidate ideas, and evidence of changing goals or preferences.
Start with one real question
After installation, open a new Codex task, call $knowledge-vault, and describe the one thing you most want to clarify.
python -X utf8 scripts/install_codex_skills.py <codex-skills-directory> --profile full --dry-run
python -X utf8 scripts/install_codex_skills.py <codex-skills-directory> --profile full
python -X utf8 scripts/install_codex_skills.py <codex-skills-directory> --profile full --verifyInitial setup begins with an interview and a preview of proposed files. The skill writes only after authorization.
Methods behind the workflows
Each skill has a source and a stated boundary. These references explain design choices; they do not establish that the complete suite has been experimentally validated.
Knowledge and provenance
W3C PROV supports source traceability. Sönke Ahrens's How to Take Smart Notes informs atomic notes and meaningful links.
Personalization and growth
ISO 9241-210 supports understanding people and context first. Deliberate-practice research and Make It Stick inform practice and retrieval evidence.
Experience and reflection
David Kolb's experiential learning cycle and Donald Schön's The Reflective Practitioner connect action, reflection, and the next experiment.
Sources and limits
The source, translation, and limits for all 11 skills are listed in the theoretical foundations.
Clear boundaries before automation
- AI inference is not recorded as fact
- One success is not stable capability
- Output does not become a source of truth
- Candidate interests are not automatic goals
- Writes require authorization
- No third-party Obsidian plugin is required
- Self-observation is not diagnosis
- Public examples are fictional
Version 0.1.3 is released under the MIT License. Automated checks cover repository contracts, links, privacy, installation, and minimum-vault creation; they are not presented as complete live Codex behavior evaluation.
Frequently asked questions
These answers also define the current limits of the project.
Is this an Obsidian plugin?
No. It is a Codex skill suite built around portable Markdown files. Obsidian Properties, Wikilinks, Bases, and dashboard conventions are available as an optional adapter.
Can I use it with an existing vault?
Yes. The bootstrap workflow inspects existing structure and migration constraints instead of forcing a fixed directory design.
Will it modify files automatically?
Important writes and initial setup are previewed first. Without write authorization, the skills provide proposals only.
Does atomic note-taking mean splitting everything into tiny files?
No. A separate atomic note is useful only when one reusable question or conclusion deserves independent retrieval. Existing knowledge should be updated first.
Can it analyze my personality?
It supports evidence-based self-observation with context, confidence, alternatives, and counterevidence. It does not diagnose or infer a fixed personality from a single event.
Has the complete system been experimentally validated?
No. Individual modules draw from established work on provenance, note-taking, deliberate practice, experiential learning, and reflection. The complete suite still needs broader user feedback and live behavior evaluation.