</>MCP Agents Market
Plugin

engram

by nagisanzenin1.4kPythonUpdated 2026-08-18

Evidence-based learning engine for Claude Code — first-principles curricula, free-recall verification with receipts, FSRS-scheduled memory, and explorable artifacts. Learn anything; keep it.

Claude CodeOpenAI CodexOpenCodeHermes AgentGoogle AntigravityOpenClawPiDeepSeek Harness

Engram is an evidence-based learning engine that integrates with Claude Code and eight other AI coding platforms, functioning as a personal tutor for developers. Unlike agent-memory plugins that enhance what the AI remembers, Engram ensures the human user retains knowledge through Socratic questioning, blind-graded free recall, and FSRS-4.5 spaced repetition scheduling. The plugin adds three commands (/learn, /review, /coach) that transform AI explanations into durable long-term memory with measurable retention tracking. All learning data stays local in plain JSON files at ~/.claude/learning/, with no cloud dependency or accounts required.

Key Features

First-principles curriculum generation that breaks topics into concept dependency graphs rather than chapter order
Socratic tutoring mode requiring active prediction and explanation-back before revealing answers
Blind assessor grading with 0/258 leniency inflation on adversarial gold set and auditable receipts for every judgment
FSRS-4.5 spaced repetition scheduler fitted to individual forgetting curves after 64+ reviews
Interactive HTML explorables with prediction gates for threshold concepts, measured per-learner for retention impact
Worked-example ladder and step-graded problem solving for procedural skills (math, git workflows, etc.)
Cross-platform state sharing: learn in Claude Code, review in OpenCode, same schedule across nine agent platforms
100% local data storage with permanent selftest blocking network code from ever entering the engine

Use Cases

  • 01Learning complex programming concepts (Rust lifetimes, transformers, Kalman filters) with retention verified at 30-90 day intervals
  • 02Mastering STEM procedures (calculus, statistics, linear algebra) through worked examples and fresh-variant practice
  • 03Retaining music theory, history, or anatomy with mnemonics for memorization-heavy content
  • 04Building durable knowledge from AI explanations instead of forgetting them days after the conversation
  • 05Tracking real competence growth with calibration metrics and retention receipts rather than self-assessment
  • 06Contributing anonymized learning data to open science studying AI-tutoring retention at month-scale horizons

Related Plugins

View more

andrej-karpathy-skills

A single CLAUDE.md file to improve Claude Code behavior, derived from Andrej Karpathy's observations on LLM coding pitfalls.

Plugin216.1k

claude-mem

Persistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More

Plugin95k

Understand-Anything

Graphs that teach > graphs that impress. Turn any code into an interactive knowledge graph you can explore, search, and ask questions about. Works with Claude Code, Codex, Cursor, Copilot, Gemini CLI, and more.

Plugin84.9k

rtk

CLI proxy that reduces LLM token consumption by 60-90% on common dev commands. Single Rust binary, zero dependencies

Plugin82.1k

engram — FAQ

What is Engram and how does it differ from agent memory plugins?+

Engram is a learning system for the human user, not the AI agent. While agent-memory plugins help the AI remember context and conversations, Engram uses the AI as a tutor to install and schedule knowledge in your own long-term memory through spaced repetition and verified recall.

How do I install Engram in Claude Code?+

Run 'claude plugin marketplace add nagisanzenin/engram' followed by 'claude plugin install engram@engram' in your terminal. After installation, use /learn, /review, and /coach commands inside Claude Code sessions. Requires Python 3 (standard library only).

Which AI coding assistants support Engram?+

Engram works with Claude Code, OpenAI Codex, OpenCode (v1 and 2.0 beta), Hermes Agent, Google Antigravity, OpenClaw, Pi, DeepSeek Harness, and ZCode. Each platform has specific installation instructions in the repository's INSTALL-*.md files.

Does Engram require API keys or a subscription?+

No API keys, accounts, or subscriptions are required. All learning data is stored locally in ~/.claude/learning/ as plain JSON files. The only network activity is the curriculum architect using web search when building new topic maps, which can be disabled for offline use.

Is Engram free and open source?+

Yes, Engram is released under the MIT license and is completely free. The code is open source on GitHub and the engine has permanent selftests preventing any network code from being added.

How does the blind grading system work?+

After each lesson, a separate AI assessor grades your free-recall answers without seeing the teaching dialogue, only the rubric and your words. This assessor has been audited on an 89-item adversarial gold set with 0/258 leniency inflations, and every grade is saved as a verifiable receipt on disk.

How do I install engram?+

Open the source repository on GitHub and follow its README. engram is a plugin — MCP Agents Market links you directly to the official repo.

Is engram free?+

engram is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.

Related searches

Knowledge & Memory PluginsProductivity & Workflow PluginsDeveloper Tools PluginsBest MCP ServersBest SkillsBest AgentsTop rated listingsClaude Code toolsBrowse all categories