Personal knowledge base: bookmarks, learning tracks, research notes, and quick captures.
| File | Purpose |
|---|---|
bookmarks.jsonl |
Saved resources — articles, tools, repos, papers |
learning.yaml |
Skills tracker and learning goals |
competitors.md |
Reference projects and landscape analysis |
research/ |
Topic deep-dives linking relevant bookmarks |
notes/ |
Quick captures and fleeting thoughts |
bookmarks.jsonl
{"id": "bk_YYYYMMDD_HHMMSS", "url": "...", "title": "...", "category": "...", "tags": ["..."], "status": "unread|read|reviewed|archived", "rating": 0, "summary": "...", "why_saved": "...", "saved": "YYYY-MM-DD"}Use one primary category per bookmark:
ai-agents— LLM tooling, agents, context engineeringarchitecture— system design, patterns, distributed systemsfrontend— UI, React, browser APIs, CSSbackend— APIs, databases, infrastructuredevops— CI/CD, containers, cloud, observabilitysecurity— auth, vulnerabilities, hardeningtooling— dev tools, editors, CLIs, productivitycareer— growth, management, communicationopen-source— OSS strategy, community, contributionresearch— papers, deep technical topics
Lowercase, singular. Reuse existing tags when possible.
Common: typescript, javascript, rust, go, python, react, node, postgres, redis, docker, kubernetes, llm, testing, performance, refactoring, api-design
Save a bookmark:
- Extract URL and title (infer from URL if not provided)
- Assign category + at least 2 tags
- Add
why_saved: why this matters, not just what it is - Set
status: unreadinitially - Check for duplicate URL before appending to
bookmarks.jsonl
Search bookmarks:
- Read
bookmarks.jsonlline by line - Filter by category or tag match
- Return title + URL + why_saved
Research a topic:
- Search bookmarks by relevant tags/category
- Summarize what's saved
- Create a note in
research/TOPIC.mdif synthesizing - Note gaps worth filling