# learn

learn is a Claude Code command that distills evaluation entries from .lihfolk/ledger.jsonl into .lihfolk/recipes.md. It extracts working patterns to raise initial generation scores.

- command · `type=command` · Claude Code, Codex, GitHub Copilot, Gemini CLI · ★ 98 repo stars
- Author: LihFolk
- License: LicenseRef-Proprietary · policy=link-only · description only; the file stays with the author
- Source: https://github.com/lihfolk-cy/v2-vault/tree/HEAD/commands/learn.md
- Updated: 2026-10-29
- Security check: automated, 2026-11-17 (text) — no signs of malicious code found.
- HTML: https://affpapa.org/en/nevibecode/commands/lihfolk-learn

## What it does

The command triggers the fifth step (Distill) of the `quality-score` skill against records in `.lihfolk/ledger.jsonl`. It targets a specified category — such as `model`, `animation`, `system`, `ui`, `world`, or `vfx` — and defaults to `all`.

For every populated category, it tallies average initial scores, revision rounds, and how often a draft hits a 9 on the first try. It isolates settings shared across entries rated 8.5 and above, criteria that fail on the first run at least 30% of the time, and positive adjustments logged twice or more.

Updates are written to `.lihfolk/recipes.md`, keeping sections under roughly 60 lines while pruning conflicting takeaways. It concludes with a trend report showing whether first-draft scores are rising. The command requires `.lihfolk/ledger.jsonl` and the `quality-score` skill.

## Install

### Claude Code (native platform)

1. **Plugin marketplace**
In the agent chat (not a shell):
```text
/plugin marketplace add lihfolk-cy/v2-vault
```
In the agent chat (not a shell):
```text
/plugin install lihfolk@lihfolk
```

2. **Manually: put the files in place**
- File: `.claude/commands/learn.md` — Call it with /learn. For all projects use ~/.claude/commands/

### Codex

**Manually: put the files in place**
- File: `~/.codex/prompts/learn.md` — Call it with /prompts:learn. Codex marks this format deprecated in favour of skills

### GitHub Copilot

**Manually: put the files in place**
- File: `.github/prompts/learn.prompt.md`

### Gemini CLI

**Manually: put the files in place**
- File: `.gemini/commands/learn.toml` — Gemini stores commands as TOML: the prompt text goes into the prompt field

Where a file is given instead of a command, copy the files from the source to that path: https://github.com/lihfolk-cy/v2-vault/tree/HEAD/commands/learn.md

## When to use it
- summarize winning defaults from assets scoring 8.5 or higher
- detect recurring first-draft failure modes above the 30% threshold
- refresh `.lihfolk/recipes.md` while discarding obsolete rules
- evaluate whether initial generation quality trends upward

## Example prompts
- /learn ui
- /learn vfx
- /learn all

## Similar
- [prompt-optimizer](https://affpapa.org/en/nevibecode/skills/getsentry-prompt-optimizer.md) — skill · 2500 installs
- [skill-optimizer](https://affpapa.org/en/nevibecode/skills/skill-optimizer.md) — skill · 1400 installs
- [senior-prompt-engineer](https://affpapa.org/en/nevibecode/skills/senior-prompt-engineer.md) — skill · 970 installs
- [skill-autoresearch](https://affpapa.org/en/nevibecode/skills/skill-autoresearch.md) — skill · 323 installs
- [autoresearch](https://affpapa.org/en/nevibecode/skills/pedronauck-autoresearch.md) — skill · 152 installs
- [prompt-optimize](https://affpapa.org/en/nevibecode/skills/prompt-optimize.md) — skill · ★ 185 repo stars

## Other versions with the same name
- [learn · worldflowai](https://affpapa.org/en/nevibecode/commands/affaan-mustafa-learn.md) — ★ 6739 repo stars
- [learn · rohitg00](https://affpapa.org/en/nevibecode/commands/learn.md) — ★ 2966 repo stars

From the same set: [lihfolk-cy/v2-vault](https://affpapa.org/en/nevibecode/by/lihfolk-cy/v2-vault.md) — 29 cards

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md
