learn

A command that turns useful lessons from a coding session into connected knowledge-graph notes. It looks for existing related notes first, then proposes patterns, failed approaches, workflows, and debugging solutions for approval.

In plain words
What is it for?
Use it to extract session learnings, save approved entries, link them to related graph notes, and focus the extraction on a supplied topic.
Why use it?
It preserves discoveries that might otherwise disappear when the session ends and reduces duplicate entries by checking existing knowledge. This makes lessons reusable across projects.

Command for Claude Code

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/peaky8linders/claude-cortex/learn
Clone the repo
git clone --depth 1 https://github.com/Peaky8linders/claude-cortex

Made for: Claude Code.

Per session 15 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 939 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 $0.00015 $0.00939
Opus 5 $0.00008 $0.00469
Sonnet 5 $0.00003 $0.00188
Haiku 4.5 $0.00002 $0.00094

Measured 2d ago against content hash 6c984bfd09c9, 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 2d ago.

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.

.claude/commands/learn.md · 127 lines

How it starts

The opening of the file, as written. The whole thing — 127 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/learn — Extract & Save Session Learnings to Knowledge Graph

You are the learning extraction agent for the Brainiac cross-project knowledge graph. Your job is to analyze the current session, propose knowledge entries, and save approved ones as graph nodes with auto-linking.

System Location

  • Engine: ~/.claude/knowledge/brainiac/
  • Graph data: ~/.claude/knowledge/graph/ (nodes.json, edges.json, embeddings.npz)
  • CLI: cd ~/.claude/knowledge && python -m brainiac <command>

If the user provided a topic hint after the command, use it to focus the search in Step 1.

Step 1: Check Existing Knowledge

Before proposing anything, search the graph for related entries:

cd ~/.claude/knowledge && python -m brainiac search "RELEVANT_TOPIC"

This prevents duplicates and shows what's already captured.

Step 2: Analyze the Session

Review the conversation and identify:

  1. Patterns discovered — Reusable approaches that worked well
  2. Anti-patterns encountered — Approaches that failed, with evidence
  3. Effective workflows — Claude Code workflows or agent configurations
  4. Solutions found — Debugging solutions for specific error classes
  5. Decisions made — Architecture decisions with rationale
  6. Hypotheses to test — Claims that emerged but aren't validated

Filter aggressively. Only propose entries that:

  • Are generalizable across projects
  • Have evidence from this session
  • Are not already in the graph (checked in Step 1)
  • Would save time if encountered again

Step 3: Propose Entries

For each proposed entry, present:

### Proposed: [type] — [name]
**Type**: pattern | antipattern | workflow | hypothesis | solution | decision
**Tags**: [relevant tags]
**Projects**: [which projects]
**Summary**: [2-3 sentences]
**Evidence**: [from this session]
**Causal links**: [if this learning was caused by or led to another entry, note it]

Ask the user which entries to save.

Step 4: Save Approved Entries to Graph

Read the full file on GitHub · 127 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. 2d ago First seen · 127 lines · 15 tokens per session scan A 6c984bfd09c9

Subscribe to this mod's changes

learn is a command published in the GitHub repository Peaky8linders/claude-cortex (11 stars, last pushed 2mo ago), licensed MIT. It adds 15 tokens to every session and 939 once invoked, about $0.0001 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 2026-08-30.