learn

A command that records useful lessons discovered during a development session, including solved problems, technical insights, patterns, and edge cases.

In plain words
What is it for?
Use it to review suggested lessons from the current work, select the ones worth keeping, and add them to the project's learning record.
Why use it?
It turns session-specific discoveries into notes you can remember and reuse later.

Command

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 commands/ankushdixit/claude-plugins/learn
Clone the repo
git clone --depth 1 https://github.com/ankushdixit/claude-plugins
Per session 6 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,176 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 $0.00006 $0.01176
Opus 5 $0.00003 $0.00588
Sonnet 5 $0.00001 $0.00235
Haiku 4.5 $0.00001 $0.00118

Measured yesterday against content hash 63810c63ffc6, 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.

solokit/commands/learn.md · 149 lines

How it starts

The opening of the file, as written. The whole thing — 149 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Learning Capture

Record insights, gotchas, and best practices discovered during development.

Step 1: Analyze Session and Generate Learning Suggestions

Review what was accomplished in the current session:

  • What code was written/changed
  • What problems were solved
  • What patterns or approaches were used
  • What technical insights were discovered
  • What gotchas or edge cases were encountered

Generate 2-3 learning suggestions based on the session work. Good learnings are:

  • Specific technical insights (not generic)
  • Actionable and memorable
  • About tools, patterns, or gotchas encountered
  • Clear and concise (1-2 sentences)

Step 2: Ask User to Select Learnings

Use AskUserQuestion with multi-select to let user choose learnings:

Question: Select Learnings from This Session

  • Question: "I've identified some potential learnings from this session. Select all that apply, or add your own:"
  • Header: "Learnings"
  • Multi-select: true
  • Options (up to 4 total):
    • Option 1: [Your generated learning suggestion 1]
    • Option 2: [Your generated learning suggestion 2]
    • Option 3: [Your generated learning suggestion 3] (if applicable)
    • Option 4: [Your generated learning suggestion 4] (if applicable)

Example Options:

  • "TypeScript enums are type-safe at compile time but add runtime overhead"
  • "Zod schemas can be inferred as TypeScript types using z.infer<>"
  • "React useCallback dependencies must include all values used inside the callback"

Step 3: For Each Selected Learning, Determine Category

For each learning the user selected (or entered), automatically suggest the most appropriate category:

Categories:

  • architecture_patterns - Design decisions, patterns used, architectural approaches
  • gotchas - Edge cases, pitfalls, bugs discovered
  • best_practices - Effective approaches, recommended patterns
  • technical_debt - Areas needing improvement, refactoring needed
  • performance_insights - Optimization learnings, performance improvements
  • security - Security-related discoveries, vulnerabilities fixed

Read the full file on GitHub · 149 lines

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 · 149 lines · 6 tokens per session scan A 63810c63ffc6

Subscribe to this mod's changes

learn is a command published in the GitHub repository ankushdixit/claude-plugins (3 stars, last pushed 7mo ago), licensed MIT. It adds 6 tokens to every session and 1,176 once invoked, about $0.0000 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 2026-08-31.