Prompt · December 31, 2025
Antigravity Tutor
A personalized, autonomous learning agent that builds a curriculum based on the 'Unknown Unknowns' in a user's reading history.
Prompt: "Antigravity" Knowledge Gap Filler
Goal: A personalized, autonomous learning agent that builds a curriculum based on the 'Unknown Unknowns' in a user's reading history.
Prompt:
"Design a background agent tailored for the 'Google Antigravity' ecosystem.
**The Problem:You read tech articles, but often skip over jargon. Over time, this creates a 'swiss cheese' knowledge base where you lack foundational concepts.
**1. Input Source:
- The agent accepts a list of URLs or text dumps (e.g., from a 'Read Later' app export).
- It also maintains a
user_profile.json:{'known_topics': ['Python', 'Basic SQL'], 'reading_level': 'Intermediate'}. **2. The Gap Analysis Engine:
- Gemini 3 Task: Scan the articles. Cross-reference mentioned concepts against the
user_profile.- Identify 'Gaps': Terms critical to the article's meaning that are likely outside the user's profile.
- Filter: Ignore trivial terms. Focus on high-value concepts (e.g., 'CRDTs', 'Paxos', 'Zero-Knowledge Proofs'). **3. The Curriculum Builder (Spaced Repetition):
- For each gap, generate a 'Micro-Lesson' card.
- Format:
- The Hook: An analogy (e.g., 'Paxos is like a group of friends trying to agree on a pizza topping...').
- The Visual: Generate a Mermaid.js flowchart syntax explaining the concept.
- The Code: A tiny, pseudo-code implementation.
- Schedule: Implement a basic SM-2 (SuperMemo-2) algorithm to schedule when this lesson should be emailed to the user (Next review: 1 day, 3 days, 7 days). **4. Delivery Mechanism:
- Generate a responsive HTML email template containing the day's lessons. **Deliverables:
gap_detector.py: The analysis logic.scheduler.py: The spaced repetition logic / database (SQLite).teacher.py: The content generation prompt."