5 Learning Capsules (5LC)

A portable, tool-agnostic system prompt that turns any AI coding assistant into an expert technology trainer. Drop it into Claude Code, Cursor, Windsurf, Aider, or any vibe coding tool, and learn a new technology in five focused 45-minute sessions.

What it is

5LC is a methodology plus a system prompt that produces a complete, hands-on study plan for any technology you want to learn. Each plan is split into five capsules:

#	Level	Focus	Outcome
1	L100	TL;DR	Know what it is and why it matters
2	L200	Intermediate	Use the common 80% of features
3	L300	Best practices	Avoid pitfalls, write production-grade code
4	Capstone	Hands-on project	Ship a working artifact in 45 min
5	Recap	Synthesis	Decision framework + roadmap for continued learning

Every session is 45 minutes. Every session ends with something you can keep — code, a query, a diagram, a decision.

Why use it

• Time-boxed. Five 45-minute slots fit into a normal week.
• Portable. Works with any AI coding assistant that can read files and follow a system prompt.
• Reproducible. Same structure every time, so capsules across topics stack into a personal learning library.
• Hands-on. Capstone in capsule 4 forces you to build, not just read.

Quick start

1. Install the prompt

Copy 5LC-PROMPT.md into your project, or reference it from your tool's system instruction file.

Claude Code — add to CLAUDE.md (project) or ~/.claude/CLAUDE.md (global):

When the user says "Create a 5LC on {topic}", automatically start the
5 Learning Capsules protocol following the structure defined in
5LC-PROMPT.md.

Cursor / Windsurf — same trigger line, in .cursorrules or .windsurfrules.

Any tool — paste the contents of 5LC-PROMPT.md directly into the system prompt or rules file.

2. Trigger the protocol

In a new conversation, say:

Create a 5LC on Apache Iceberg

The assistant will ask three onboarding questions:

1. Topic confirmation
2. Optional documentation links
3. Session cadence (how many per week)

Answer them, and the assistant generates the full folder structure and (optionally) schedules calendar events.

3. Run the capsules

Open 5-learning-capsules/{topic}/Capsule_1/agenda.md and start. Each capsule is self-contained.

Output structure

5-learning-capsules/{topic}/
├── README.md                       # Plan overview
├── {topic}-cheatsheet.md           # Quick reference
└── Capsule_{1..5}/
    ├── agenda.md                   # Session timing
    ├── content.md                  # Guided study material
    ├── exercises.md                # Hands-on steps
    └── resources.md                # Links and references

Supported tools

5LC has been tested with:

• Claude Code (CLI + IDE extensions)
• Cursor
• Windsurf
• Aider
• Continue.dev

Any tool that supports a project-level system prompt and can write files will work. See 5LC-PROMPT.md for the recognized instruction-file names per tool.

Example topics

Good fits for the 5LC format:

• A specific technology — Apache Iceberg, DuckDB, Polars, Temporal
• A framework — FastAPI, Next.js App Router, LangGraph
• A platform feature — Databricks Lakebase, Snowflake Cortex, AWS Bedrock
• A concept — vector search, MLOps, event sourcing

Less good fits: huge ecosystems ("learn AWS"), open-ended skills ("become a better manager"). Scope down first.

Customization

The prompt is intentionally short and editable. Common tweaks:

• Language. Replace English with your preferred language for the generated content.
• Session length. Change "45 min" to whatever fits your schedule. Adjust capstone scope accordingly.
• Folder naming. Switch Capsule_ to Sesion_, Modulo_, etc.
• Calendar. Remove the calendar block if you don't want events scheduled.

License

MIT — see LICENSE.

Contributing

Open an issue or PR. Improvements to the prompt, additional tool integrations, and example capsules from real topics are all welcome.