learn

A command for reviewing what happened during a coding session or answering a question about the work recorded in its session log. It can also check saved learning notes when relevant.

In plain words
What is it for?
Use it to recap completed work, recall a topic from the session, or answer questions about how something was built.
Why use it?
It helps recover context without manually rereading the whole session history. It also highlights when the session revisits a topic marked for review.

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/gkaria/vibe-learn/learn
Clone the repo
git clone --depth 1 https://github.com/gkaria/vibe-learn

Made for: Claude Code.

Per session 21 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,377 The whole file, excluding the scripts and references it only reads on demand.
Security scan B 1 finding. 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 $0.00021 $0.01377
Opus 5 $0.00010 $0.00688
Sonnet 5 $0.00004 $0.00275
Haiku 4.5 $0.00002 $0.00138

Measured 2d ago against content hash 19cbdecfdd5b, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade B, and why

learn scanned grade B with 1 finding 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.

Reads agent configuration directoriesmediumAgent snooping

.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.

**Knowledge ledger check (all modes):** if `.vibe-learn/knowledge.json` exists, run the knowledge helper's `due` command via Bash (`bash ~/.vibe-learn/scripts/knowledge.sh due`; if that path doesn't exist, use the `scrip
.claude/commands/learn.md · 111 lines

How it starts

The opening of the file, as written. The whole thing — 111 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Read .vibe-learn/session-log.jsonl to understand what has happened in this session.

Knowledge ledger check (all modes): if .vibe-learn/knowledge.json exists, run the knowledge helper's due command via Bash (bash ~/.vibe-learn/scripts/knowledge.sh due; if that path doesn't exist, use the scripts/knowledge.sh next to the bootstrap.sh your vibe-learn hooks point at in .claude/settings.local.json or ~/.claude/settings.json, or bash scripts/knowledge.sh due when working in the vibe-learn repo itself). If a due concept was also touched in this session's log, open your response with a single heads-up line, e.g. "Heads up: this session touched JWT refresh tokens again — you marked that shaky on July 11. Want a quick recap first, or /quiz review?" At most one such line, and never let it block or replace the actual answer. If the ledger or helper is missing, skip this silently.

Parse $ARGUMENTS to determine the mode:


Mode: obsidian:recall <topic>

If $ARGUMENTS starts with obsidian:recall:

Extract the topic (everything after obsidian:recall). If no topic is given, ask the user what they want to recall.

  1. Load the Obsidian config: check .vibe-learn/obsidian.json first, then ~/.vibe-learn/obsidian.json. If neither exists, run the config setup flow (ask vault path, subfolder, global vs project save location) before proceeding — same flow as the obsidian mode below.
  2. Construct the search path: <vault_path>/<subfolder>/
  3. Search that directory for markdown files matching the topic — search across filenames, headings (#), body text, and frontmatter tags using the Grep tool.
  4. Read up to 5 of the most relevant matching files.
  5. Synthesize a summary covering:
    • Which sessions touched this topic and when (use the date frontmatter field)
    • Key decisions and patterns related to the topic
    • Any unchecked - [ ] items in "Things to Study" sections on this topic
    • Connections or progressions across sessions (e.g., "You first set this up on March 2nd, then extended it on April 5th")
  6. Present the summary to the user. Do not write any file.

Read the full file on GitHub · 111 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 · 111 lines · 21 tokens per session scan B 19cbdecfdd5b

Subscribe to this mod's changes

learn is a command published in the GitHub repository gkaria/vibe-learn (54 stars, last pushed 6d ago), licensed MIT. It adds 21 tokens to every session and 1,377 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.