hammer

hammer

Personal agent-skills for learning. The agent plays tutor, you play learner — it refuses to just hand you answers, keeps a persistent profile of what you know and what you've stumbled on, and spaces review of past material across sessions.

Currently ships one skill: teach. The repo is named for Richard Hamming, whose Learning to Learn is the spine of the approach.

Install

The repo is in the vercel-labs/skills format and also carries a Claude Code plugin manifest, so you can install it via whichever path you prefer.

Via npx skills (any supported agent)

# install the teach skill globally (available to all projects)
npx skills add SebastianElvis/hammer -g

# or scoped to the current project
npx skills add SebastianElvis/hammer

# target a specific agent (claude-code, codex, cursor, opencode, …)
npx skills add SebastianElvis/hammer -g -a claude-code

Supported agents are listed in the vercel-labs/skills README. The CLI auto-detects what you have installed.

As a Claude Code plugin

This repo doubles as a Claude Code plugin marketplace. Add the marketplace, then install the hammer plugin:

/plugin marketplace add SebastianElvis/hammer
/plugin install hammer@hammer

After installing, run /reload-plugins to activate it.

Manually

Clone the repo and symlink skills/teach/ into your agent's skills directory (e.g., ~/.claude/skills/teach for Claude Code).

Usage

Once installed, just talk to the agent the way you already do. The skill triggers automatically when you ask to learn something rather than get something done:

• "teach me SQL window functions"
• "let's learn Rust ownership — don't just explain, walk me through it"
• "quiz me on what we covered last time"
• "I want to understand how a Bloom filter works"

It will not fire on debugging help, one-shot factual lookups, or task execution — those are not tutoring. See skills/teach/SKILL.md for the full trigger definition.

Where your progress lives

The skill stores your learner profile outside the skill directory, so reinstalling or updating the skill never touches your data.

• Default location: ./.teach/ — the .teach/ folder in the agent's current working directory. Learner state is scoped per project; .teach/ is gitignored so it does not get committed.
• Override: set TEACH_HOME to any path — useful for a cross-project profile (TEACH_HOME=~/.teach) or per-topic folders (TEACH_HOME=~/.teach/spanish).

The folder contains:

./.teach/
├── learner.md        # your profile — currently studying, goal, known solid, shaky, misconceptions, calibration notes
├── review.md         # spaced-review queue (new / learning / mastered buckets)
├── syllabus.md       # ordered arc for a multi-session topic (created only when needed)
└── sessions/         # one markdown transcript per day

Everything is plain markdown. You can read it, edit it, back it up, sync it via git/Dropbox — it's just text.

How a session flows

The lifecycle runs linearly: calibration → prepare-syllabus → teach loop. Each upstream phase is either run or skipped based on state; teaching is always last.

┌────────┐   ┌─────────────┐   ┌──────────────────┐   ┌──────────────────┐   ┌────────┐
│ START  │   │ CALIBRATION │   │ PREPARE-SYLLABUS │   │   TEACH LOOP     │   │  END   │
│        │──▶│             │──▶│                  │──▶│                  │──▶│        │
│ read   │   │  skip if    │   │   skip if no     │   │ drill ⇄ socratic │   │ update │
│ state  │   │  calibrated │   │   arc needed     │   │        ⇅         │   │ state  │
│        │   │             │   │                  │   │     feynman      │   │        │
└────────┘   └─────────────┘   └──────────────────┘   └──────────────────┘   └────────┘

Invariant inside the teach loop: answer-protection on the current target question persists across mode switches — a mode switch is not a way out of a stuck Socratic question. See references/refusal-rules.md.

Step by step: