learn

learn is a skill for Claude Code, Codex from morodomi/dev-crew. It costs 38 tokens per session (867 once invoked), scanned A, original, MIT.

A learning workflow that examines recent project work and records repeated patterns as reusable instincts. It uses cycle documents, Git history, changed files, and logged observations.

In plain words
What is it for?
Use it to review recent work, identify patterns such as error fixes or workflow preferences, and pass them to an observer for storage.
Why use it?
It helps preserve useful lessons from past sessions instead of leaving them in a developer's memory or scattered across history.

Skill for Claude CodeCodex

Part of the dev-crew plugin — 29 skills, 41 agents, 3 hooks shipped together

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/morodomi/dev-crew/learn
Any agent
npx skills add morodomi/dev-crew --skill learn
Clone the repo
git clone --depth 1 https://github.com/morodomi/dev-crew

Made for: Claude Code, Codex.

Or install dev-crew, the plugin that ships this one along with the rest of its 29 skills, 41 agents, 3 hooks.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for learn

README.md
[![agentmods](https://agentmods.dev/badge/skills/morodomi/dev-crew/learn.svg)](https://agentmods.dev/skills/morodomi/dev-crew/learn)
Your own site
<a href="https://agentmods.dev/skills/morodomi/dev-crew/learn"><img src="https://agentmods.dev/badge/skills/morodomi/dev-crew/learn.svg" alt="Measured on agentmods" height="20"></a>
Per session 38 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 867 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00038 $0.00867
Opus 5 $0.00019 $0.00434
Sonnet 5 $0.00008 $0.00173
Haiku 4.5 $0.00004 $0.00087

Measured yesterday against content hash 21f1eb120a3f, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

learn scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/learn/SKILL.md · 95 lines

What it actually says

Learn - セッションパターン抽出

セッションの作業パターンを検出し、instinct (構造化パターン) として蓄積する。

前提

  • プロジェクトディレクトリ内で実行すること
  • Git 管理下であること (git log / git diff を使用)

実行手順

Step 1: 初回チェック (Empty State)

蓄積先ディレクトリを確認。存在しない場合は自動作成しオンボーディングメッセージを表示:

dev-crew learn へようこそ。
learn はセッションの作業パターンを記録し、evolve でスキルに進化させます。
初回セットアップを実行しました。
mkdir -p ${CLAUDE_PLUGIN_DATA}/instincts/

Step 2: 入力ソース収集 + TF-IDF サマリ計算

以下の情報を収集する:

ソース コマンド
Cycle doc ls -t docs/cycles/*.md 2>/dev/null | head -1
変更履歴 git log --oneline -20
変更ファイル git diff --name-only HEAD~5..HEAD
observations cat ${CLAUDE_PLUGIN_DATA}/observations/log.jsonl 2>/dev/null
TF-IDF サマリ計算

scripts/tfidf-summary.sh で observations (log.jsonl) からセッション横断の TF-IDF スコアを算出し、tfidf_summary として observer に渡す。

TFIDF_JSON="$(bash scripts/tfidf-summary.sh)"

空配列 [] が返った場合 (ブートストラップ期間 or データなし)、observer に tfidf_summary を渡さない。

詳細: reference.md

Step 3: ユーザー補足情報

AskUserQuestion で確認:

  • 「このセッションで気づいたパターンはありますか?」
  • 選択肢: エラー解決 / ワークフロー改善 / ツール選好 / 特になし

Step 4: observer エージェントに委譲

Task tool で observer を起動し、収集した入力ソースを渡す:

Task(subagent_type: "dev-crew:observer", prompt: "...")

Step 5: 品質フィルタ + 保存

observer の出力から instinct を選別:

confidence 判定
>= 0.5 保存 (${CLAUDE_PLUGIN_DATA}/instincts/ に JSONL 追記)
< 0.5 破棄 (理由付きで報告)

confidence >= 0.8 の instinct は MEMORY.md 昇格候補としてユーザーに提示する。棲み分けルール詳細: reference.md

Step 6: 結果報告

N 件のパターンを保存しました。
現在の蓄積: X 件。evolve 可能まであと Y 件。

保存成功後、タイムスタンプを更新:

date -u +"%Y-%m-%dT%H:%M:%SZ" > ${CLAUDE_PLUGIN_DATA}/observations/.last-learn-timestamp

Reference

詳細: reference.md

Files

What ships with it

1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 95 lines · 38 tokens per session scan A 21f1eb120a3f

Subscribe to this mod's changes

learn is a skill published in the GitHub repository morodomi/dev-crew (1 stars, last pushed yesterday), licensed MIT. It adds 38 tokens to every session and 867 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.