vscode-knowledge-manager README

A tool to help you manage your knowledge in vscode.

Features

Knowledge review remind it reminds you to review your knowledge according to forgetting curve.
Review Log System Automatically records all review activities with Git-friendly format.
- Tracks rating, stability changes, and review timing
- JSONL format for easy Git merging and conflict resolution
- Cross-device sync via Git repository
- Per-file storage in .assets directories
- Smart file rename synchronization
- Enables future parameter optimization based on actual review history
Customizable FSRS Parameters Configure spaced repetition algorithm parameters to match your learning style.
- Choose from presets: Aggressive, Balanced, or Conservative
- Or customize: retention rate, maximum interval, and more
- Auto-recalculation: When changing parameters, option to recalculate all existing card schedules
- Auto-optimization: Analyze your review history to calculate personalized parameters optimized for your learning patterns
AI Chat Participant (@km) Chat with AI to learn concepts and generate knowledge questions.
- /learn <knowledge> - Get detailed explanations of concepts
- /generate_questions <file> - Auto-generate practice questions from markdown files
- /check_grammar <file> - Check English grammar and spelling errors

FSRS Configuration

Quick Start - Using Presets (Recommended)

Open settings (Ctrl+, or Cmd+,), search for Knowledge Manager, find FSRS Preset option:

Balanced (Default) - For most learning scenarios
- Target retention: 90%
- Maximum interval: 100 years
Aggressive - High-frequency reviews
- Target retention: 95%
- Maximum interval: 180 days
- Best for: Important but hard-to-remember content, exam preparation
Conservative - Low-frequency reviews
- Target retention: 85%
- Maximum interval: 100 years
- Best for: Easy-to-remember content, long-term knowledge retention

Custom Configuration

Choose Custom preset and manually adjust these parameters:

1. Request Retention (Target retention rate)

Default: 0.9 (90%)
Range: 0.7 - 0.99
Description: Probability you want to remember at next review
Suggested values:
- Exam preparation: 0.95
- Daily learning: 0.9
- Long-term retention: 0.85

2. Maximum Interval (Maximum days between reviews)

Default: 36500 (~100 years)
Range: 1 - 36500
Suggested values:
- Short-term goals (exams): 180
- Medium-term goals (work skills): 365
- Long-term goals (lifelong learning): 36500

3. Enable Fuzz

Default: true
Description: Adds randomness to review times to avoid too many cards on same day
Recommendation: Keep enabled

4. Enable Short Term

Default: false
Description: Uses shorter intervals for newly learned content
Recommendation: Usually keep disabled unless doing intensive short-term learning

Configuration Examples

Scenario 1: Exam Preparation

{
  "knowledgeManager.fsrs.preset": "aggressive"
}

Scenario 2: Learning Programming Language

{
  "knowledgeManager.fsrs.preset": "balanced"
}

Scenario 3: Maintaining Mastered Knowledge

{
  "knowledgeManager.fsrs.preset": "conservative"
}

Parameter Change Handling

Automatic Recalculation Prompt

When you modify FSRS parameters, the system will automatically ask if you want to recalculate all card due dates:

Recalculate All: Recalculates all existing cards with new parameters
- ✅ Best for switching learning strategies (conservative → aggressive)
- ✅ Ensures all cards follow unified strategy
- ⚠️ Will change existing review schedules
Apply to New Reviews Only: Keeps existing card schedules unchanged
- ✅ Best for fine-tuning parameters
- ✅ Won't disrupt current review rhythm
- ⚠️ Old and new cards may use different strategies

Manual Recalculation

You can manually trigger recalculation anytime:

Open command palette (Ctrl+Shift+P / Cmd+Shift+P)
Search for "Knowledge Manager: Recalculate All Card Due Dates"
Confirm the operation

Note: Recalculation preserves your review count and lapse count. Only reviewed cards (non-New state) will be recalculated.

When to Adjust Parameters?

Should adjust:

✅ Review intervals too long, often forget → Increase requestRetention
✅ Reviews too frequent, too burdensome → Decrease requestRetention
✅ Preparing for short-term exam → Use aggressive preset or lower maximumInterval
✅ Different projects have different needs → Use workspace settings

No need to adjust:

❌ Just started using → Observe with default settings first
❌ Insufficient review data (< 100 reviews) → Wait to accumulate more data

Troubleshooting

Problem: Parameters changed but not taking effect

Solution: Confirm you chose "Recalculate All" in the parameter change prompt

Problem: Not sure which preset to use

Solution: Start with balanced, adjust after 1-2 weeks based on experience

Problem: After switching to aggressive, cards still show due far in future

Solution: Old conservative strategy dates are still in effect. Use "Recalculate All Card Due Dates" command to apply new strategy to all cards

Chat Participant Usage

@km /learn

Analyze and explain knowledge concepts in detail:

@km /learn FSRS algorithm
@km /learn spaced repetition

@km /generate_questions

Generate ::: knowledge blocks from markdown content:

@km /generate_questions

Usage:

Open a markdown file and run @km /generate_questions to generate questions from current file
Or attach a file: @km /generate_questions #file
Or specify path: @km /generate_questions ./docs/tutorial.md

The generated questions are ready to use with FSRS tracking for spaced repetition learning.

Requirements

Markdown file front matters markdown file with front matters
- reviewTime: next review time of knowledge
- reviewCount reviewed count of knowledge
GitHub Copilot (required for @km chat participant)
- Active GitHub Copilot subscription
- Copilot extension installed and enabled

Extension Settings

None

Known Issues