cass_memory_system
bydicklesworthstone· 3 MCP servers
Procedural memory for AI coding agents: transforms scattered session history into persistent, cross-agent memory so every agent learns from every other
How to install cass_memory_system?
dicklesworthstone/cass_memory_system
$git clone https://github.com/dicklesworthstone/cass_memory_systemInstalls into the current project.
›Prefer a prompt? Paste this to your agent
Install cass_memory_system by running `git clone https://github.com/dicklesworthstone/cass_memory_system`, then use it for the current task and follow its documentation at https://github.com/dicklesworthstone/cass_memory_system.
README.md
| 1 | # cass-memory |
| 2 | |
| 3 | <div align="center"> |
| 4 | <img src="cm_illustration.webp" alt="cass-memory - Procedural memory for AI coding agents"> |
| 5 | </div> |
| 6 | |
| 7 |  |
| 8 |  |
| 9 |  |
| 10 |  |
| 11 | |
| 12 | **Procedural memory for AI coding agents.** |
| 13 | Transforms scattered agent sessions into persistent, cross-agent memory—so every agent learns from every other agent's experience. |
| 14 | |
| 15 | <div align="center"> |
| 16 | |
| 17 | **One-liner install (Linux/macOS):** |
| 18 | |
| 19 | ```bash |
| 20 | curl -fsSL "https://raw.githubusercontent.com/Dicklesworthstone/cass_memory_system/main/install.sh?$(date +%s)" \ |
| 21 | | bash -s -- --easy-mode --verify |
| 22 | ``` |
| 23 | |
| 24 | **Or via package managers:** |
| 25 | |
| 26 | ```bash |
| 27 | # macOS/Linux (Homebrew) |
| 28 | brew install dicklesworthstone/tap/cm |
| 29 | |
| 30 | # Windows (Scoop) |
| 31 | scoop bucket add dicklesworthstone https://github.com/Dicklesworthstone/scoop-bucket |
| 32 | scoop install dicklesworthstone/cm |
| 33 | ``` |
| 34 | |
| 35 | </div> |
| 36 | |
| 37 | --- |
| 38 | |
| 39 | ## 🤖 Agent Quickstart (JSON) |
| 40 | |
| 41 | **Always use `--json` in agent contexts.** stdout = data, stderr = diagnostics, exit 0 = success. |
| 42 | |
| 43 | ```bash |
| 44 | # 1) Get task-specific memory before you start |
| 45 | cm context "implement auth rate limiting" --json |
| 46 | |
| 47 | # 2) See the minimum viable workflow |
| 48 | cm quickstart --json |
| 49 | |
| 50 | # 3) Build the playbook (memory onboarding) |
| 51 | cm onboard status --json |
| 52 | cm onboard sample --fill-gaps --json |
| 53 | cm onboard read /path/to/session.jsonl --template --json |
| 54 | cm onboard mark-done /path/to/session.jsonl |
| 55 | ``` |
| 56 | |
| 57 | ## Table of Contents |
| 58 | |
| 59 | - [Why This Exists](#-why-this-exists) |
| 60 | - [How It Works](#-how-it-works) |
| 61 | - [Key Features](#-key-features) |
| 62 | - [For AI Agents](#-for-ai-agents-the-most-important-section) |
| 63 | - [Installation](#-installation) |
| 64 | - [CLI Reference](#-cli-reference) |
| 65 | - [The ACE Pipeline](#-the-ace-pipeline) |
| 66 | - [Data Models](#-data-models) |
| 67 | - [Scoring Algorithm](#-scoring-algorithm) |
| 68 | - [Configuration](#-configuration) |
| 69 | - [MCP Server](#-mcp-server) |
| 70 | - [Architecture & Engineering](#-architecture--engineering) |
| 71 | - [Deep Dive: Core Algorithms](#-deep-dive-core-algorithms) |
| 72 | - [Privacy & Security](#-privacy--security) |
| 73 | - [Trauma Guard: Safety System](#-trauma-guard-safety-system) |
| 74 | - [Performance Characteristics](#-performance-characteristics) |
| 75 | - [Starter Playbooks](#-starter-playbooks) |
| 76 | - [Extensibility](#-extensibility-adding-new-components) |
| 77 | - [Troubleshooting](#-troubleshooting) |
| 78 | - [Design Philosophy](#-design-philosophy) |
| 79 | - [Comparison with Alternatives](#-comparison-with-alternatives) |
| 80 | - [Roadmap](#-roadmap) |
| 81 | |
| 82 | --- |
| 83 | |
| 84 | ## 💡 Why This Exists |
| 85 | |
| 86 | ### The Problem |
| 87 | |
| 88 | AI coding agents accumulate valuable knowledge through sessions: debugging strategies, code patterns, user preferences, project-specific insights. But this knowledge is: |
| 89 | |
| 90 | 1. **Trapped in sessions** — Each session ends, context is lost forever |
| 91 | 2. **Agent-specific** — Claude Code doesn't know what Cursor learned yesterday |
| 92 | 3. **Unstructured** — Raw conversation logs aren't actionable as guidance |
| 93 | 4. **Subject to collapse** — Naive summarization loses critical nuances and details |
| 94 | |
| 95 | You've solved authentication bugs three times this month across different agents. Each time, you started from scratch because the knowledge from previous sessions was inaccessible. |
| 96 | |
| 97 | ### The Solution |
| 98 | |
| 99 | `cass-memory` implements a **three-layer cognitive architecture** that transforms raw session logs into actionable, confidence-tracked rules: |
| 100 | |
| 101 | | Layer | Role | Implementation | |
| 102 | |-------|------|----------------| |
| 103 | | **Episodic Memory** | Raw ground truth from all agents | `cass` search engine | |
| 104 | | **Working Memory** | Structured session summaries | Diary entries | |
| 105 | | **Procedural Memory** | Distilled rules with tracking | Playbook bullets | |
| 106 | |
| 107 | This mirrors how human expertise develops: raw experiences (episodic) are consolidated into structured memories (working), which eventually become automatic knowledge (procedural). |
| 108 | |
| 109 | ### Who Benefits |
| 110 | |
| 111 | - **AI Agents**: Get relevant rules and historic |
Related
6 picksType
- langchain-mcp-adaptersMake Anthropic Model Context Protocol (MCP) tools compatible with LangChain and LangGraph agents.Jul 2026
- mcporterTypeScript runtime and CLI for connecting to configured Model Context Protocol servers.Jul 2026
- mcp-proxyA MCP server which proxies requests to a remote MCP server over streamable HTTP or SSE.Jul 2026
- mcp-atlassianThe Model Context Protocol (MCP) Atlassian integration is an open-source implementation that bridges Atlassian products (Jira and Confluence) with AI language models following Anthropic's MCP specification.Jul 2026
- open-webuiOpen WebUIJul 2026
- rulesyncUnified AI rules management CLI tool that generates configuration files for various AI development toolsJul 2026
