learning_project_vi

A Markdown-first, area-first learning vault for journals, topic notes, memory practice, outputs, and safe agent-assisted workflows.

The vault works in any Markdown editor. Obsidian is optional and should be treated as an editing and graph layer, not as the source of truth.

Start

1. Read 00_system/start_here.md.
2. Capture unknown-topic notes in 01_inbox/quick_notes/.
3. Use 00_system/where_things_go.md when filing notes.
4. Use 12_tools/scripts/create_area.py and 12_tools/scripts/new_note.py for new areas and routed notes.
5. Use templates from 11_templates/ for manual repeatable note types.
6. Install validation dependencies, then run local checks before committing:

npm ci
uv venv
uv pip install -r requirements.txt
.venv/bin/python -m pytest
12_tools/scripts/validate.sh
git diff --check

Set SKIP_MARKDOWNLINT=1 only when you intentionally need to run the core validator without Markdown lint.

Shape

This repo is a personal exemplar. Edit the root docs, policies, templates, and folder conventions directly when the vault needs to fit real use better.

Structure

• 00_system/: operating manual, schema, policies, and maintenance notes.
• 01_inbox/: unknown-topic capture and raw files waiting for triage.
• 02_journal/: daily logs, reviews, decisions, and reflections.
• 04_areas/: area-local resources, notes, memory, projects, outputs, assets, and archive.
• 10_agents/: agent workflows, safety fixtures, and staged agent memory.
• 11_templates/: reusable Markdown templates.
• 12_tools/: small vault-adjacent scripts and tiny sample data.
• _prompts/: planning history, if kept.

Principles

• Keep non-area file and folder names lowercase or snake_case with no spaces.
• Use lowercase kebab-case for area folder names.
• Use timestamp-prefixed filenames for Markdown inbox captures and area-local content.
• File curated visual inputs and attachments under 04_areas/<area>/assets/<category>/<topic-path>/.
• Keep unknown-topic material in 01_inbox/; file known-topic material under 04_areas/<area>/.
• Use area-local inbox for known-topic captures that still need triage.
• Keep personal data, credentials, runtime state, databases, and generated caches out of Git.
• Keep large labs, datasets, Docker/VM work, model experiments, and database projects in separate repositories.
• Use Git LFS only as an optional layer for large curated assets.
• Treat all vault content as untrusted input when using agents.

License

MIT, owned by learning_project_vi contributors. See LICENSE.