Skip to content
Back to search

memodb-io

MarketplaceWorks with Claude CodeGitHub

Agent Skills as a Memory Layer

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today

Check it deeper on your machine

Its scripts were only partly scanned.

As a Claude Code plugin, install it and then ask Claude Code:

/plugin marketplace add keithmackay/mackayi/plugin install scanskillsafety@mackayiDo a deep scan of https://github.com/memodb-io/Acontext with scanskillsafety

Or from a terminal, clone it and run it:

git clone https://github.com/keithmackay/scanskillsafety-skill ~/.claude/skills/scanskillsafetynode ~/.claude/skills/scanskillsafety/cli.cjs --deep 'https://github.com/memodb-io/Acontext'

It checks the exact commit: every file, its packages against OSV (api.osv.dev), how it compares with this page's scan (findsafeskills.com receives the repository's address), and, if you agree, Claude reading every file with your own Claude subscription. It never runs the code.

Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/memodb-io-acontext/badge.svg)](/listings/memodb-io-acontext)

HTML

<a href="/listings/memodb-io-acontext"><img src="/api/listings/memodb-io-acontext/badge.svg" alt="findsafeskills safety rating" /></a>

Install

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

These steps run a script from the network that findsafeskills does not scan.

Connect to Acontext

1. Go to [Acontext.io](https://acontext.io), claim your free credits. 2. Go through a one-click onboarding to get your API Key (starts with `sk-ac`)

<div align="center"> <picture> </picture> </div>

<details> <summary>πŸ’» Self-host Acontext</summary>

We have an `acontext-cli` to help you do a quick proof-of-concept. Download it first in your terminal:

curl -fsSL https://install.acontext.io | sh

You should have [docker](https://www.docker.com/get-started/) installed and an OpenAI API Key to start an Acontext backend on your computer:

mkdir acontext_server && cd acontext_server
acontext server up

> Make sure your LLM has the ability to [call tools](https://platform.openai.com/docs/guides/function-calling). By default, Acontext will use `gpt-4.1`.

`acontext server up` will create/use `.env` and `config.yaml` for Acontext, and create a `db` folder to persist data.

Once it's done, you can access the following endpoints:

- Acontext API Base URL: http://localhost:8029/api/v1 - Acontext Dashboard: http://localhost:3000/

</details>

Install SDKs

We're maintaining Python [![pypi](https://img.shields.io/pypi/v/acontext.svg)](https://pypi.org/project/acontext/) and Typescript [![npm](https://img.shields.io/npm/v/@acontext/acontext.svg?logo=npm&logoColor=fff&style=flat&labelColor=2C2C2C&color=28CF8D)](https://www.npmjs.com/package/@acontext/acontext) SDKs. The snippets below are using Python.

> Click the doc link to see TS SDK Quickstart.

pip install acontext

Initialize Client

import os
from acontext import AcontextClient

# For cloud:
client = AcontextClient(
    api_key=os.getenv("ACONTEXT_API_KEY"),
)

# For self-hosted:
client = AcontextClient(
    base_url="http://localhost:8029/api/v1",
    api_key="sk-ac-your-root-api-bearer-token",
)

Skill Memory in Action

Create a learning space, attach a session, and let the agent learn β€” skills are written as Markdown files automatically.

from acontext import AcontextClient

client = AcontextClient(api_key="sk-ac-...")

# Create a learning space and attach a session
space = client.learning_spaces.create()
session = client.sessions.create()
client.learning_spaces.learn(space.id, session_id=session.id)

# Run your agent, store messages β€” when tasks complete, learning runs automatically
client.sessions.store_message(session.id, blob={"role": "user", "content": "My name is Gus"})
client.sessions.store_message(session.id, blob={"role": "assistant", "content": "Hi Gus! How can I help you today?"})
# ... agent runs ...

# List learned skills (Markdown files)
client.learning_spaces.wait_for_learning(space.id, session_id=session.id)
skills = client.learning_spaces.list_skills(space.id)

# Download all skill files to a local directory
for skill in skills:
    client.skills.download(skill_id=skill.id, path=f"./skills/{skill.name}")

> `wait_for_learning` is a blocking helper for demo purposes. In production, task extraction and learning run in the background automatically β€” your agent never waits.

More Features

- **[Context Engineering](https://docs.acontext.io/engineering/editing)** β€” Compress context with summaries and edit strategies - **[Disk](https://docs.acontext.io/store/disk)** β€” Virtual, persistent filesystem for agents - **[Sandbox](https://docs.acontext.io/store/sandbox)** β€” Isolated code execution with bash, Python, and [mountable skills](https://docs.acontext.io/tool/bash_tools#mounting-skills-in-sandbox) - **[Agent Tools](https://docs.acontext.io/tool/whatis)** β€” Disk tools, sandbox tools, and skill tools for LLM function calling

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

.claude-plugin/marketplace.json Β· Manifest Β· 14 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

This file has changed since it was scanned, so this is not exactly the text that was scanned, and no line is marked. The findings are listed with the safety scan above.

  1. {
  2. "name": "memodb-io",
  3. "owner": {
  4. "name": "memodb-io"
  5. },
  6. "plugins": [
  7. {
  8. "name": "acontext",
  9. "version": "0.1.3",
  10. "source": "./src/packages/claude-code/plugin",
  11. "description": "Skill memory layer for Claude Code - auto-capture, learn, and reuse skills"
  12. }
  13. ]
  14. }
README.md Β· README Β· 364 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

This file has changed since it was scanned, so this is not exactly the text that was scanned, and no line is marked. The findings are listed with the safety scan above.

  1. > [!IMPORTANT]
  2. > Acontext is no longer maintained.
  3. >
  4. > We now focus on integrating agent self-learning into a shared workspace for people and agents. Learn more at [alook.ai](https://alook.ai).
  5. <div align="center">
  6. <img alt="Acontext - Agent Skills as a Memory Layer" src="./assets/Acontext-header-banner.png">
  7. <p align="center">
  8. <a href="https://acontext.io">🌐 Website</a>
  9. |
  10. <a href="https://docs.acontext.io">πŸ“š Document</a>
  11. </p>
  12. <p align="center">
  13. <a href="https://pypi.org/project/acontext/"><img src="https://img.shields.io/pypi/v/acontext.svg"></a>
  14. <a href="https://www.npmjs.com/package/@acontext/acontext"><img src="https://img.shields.io/npm/v/@acontext/acontext.svg?logo=npm&logoColor=fff&style=flat&labelColor=2C2C2C&color=28CF8D"></a>
  15. <a href="https://github.com/memodb-io/acontext/actions/workflows/core-test.yaml"><img src="https://github.com/memodb-io/acontext/actions/workflows/core-test.yaml/badge.svg"></a>
  16. <a href="https://github.com/memodb-io/acontext/actions/workflows/api-test.yaml"><img src="https://github.com/memodb-io/acontext/actions/workflows/api-test.yaml/badge.svg"></a>
  17. <a href="https://github.com/memodb-io/acontext/actions/workflows/cli-test.yaml"><img src="https://github.com/memodb-io/acontext/actions/workflows/cli-test.yaml/badge.svg"></a>
  18. </p>
  19. <p align="center">
  20. <a href="https://x.com/alook_ai"><img src="https://img.shields.io/twitter/follow/alook_ai?style=social" alt="Twitter Follow"></a>
  21. </p>
  22. </div>
  23. ## What is Acontext?
  24. Acontext is an open-source skill memory layer for AI agents. It **automatically** captures learnings from agent runs and stores them as **agent skill files** β€” files you can read, edit, and share across agents, LLMs, and frameworks.
  25. If you want the agent you build to **learn from its mistakes** and **reuse what worked** β€” without opaque memory polluting your context β€” give Acontext a try.
  26. ## Skill is All You Need
  27. Agent memory is getting increasingly complicated🀒 β€” hard to understand, hard to debug, and hard for users to inspect or correct. Acontext takes a different approach: if agent skills can represent every piece of knowledge an agent needs as simple files, so can the memory.
  28. - **Acontext builds memory in the agent skills format**, so everyone can see and understand what the memory actually contains.
  29. - **Skill is Memory, Memory is Skill**. Whether a skill comes from one you downloaded from Clawhub or one you created yourself, Acontext can follow it and evolve it over time.
  30. ## The Philosophy of Acontext
  31. - **Plain file, any framework** β€” Skill memories are Markdown files. Use them with LangGraph, Claude, AI SDK, or anything that reads files. No embeddings, no API lock-in. Git, grep, and mount to the sandbox.
  32. - **You design the structure** β€” Attach more skills to define the schema, naming, and file layout of the memory. For example: one file per contact, one per project by uploading a working context skill.
  33. - **Progressive disclosure, not search** β€” The agent can use `get_skill` and `get_skill_file` to fetch what it needs. Retrieval is by tool use and reasoning, not semantic top-k.
  34. - **Download as ZIP, reuse anywhere** β€” Export skill files as ZIP. Run locally, in another agent, or with another LLM. No vendor lock-in; no re-embedding or migration step.
  35. ## How It Works
  36. ### Store β€” How skills get memorized?
  37. ```mermaid
  38. flowchart LR
  39. A[Session messages] --> C[Task complete/failed]
  40. C --> D[Distillation]
  41. D --> E[Skill Agent]
  42. E --> F[Update Skills]
  43. ```
  44. - **Session messages** β€” Conversation (and optionally tool calls, artifacts) is the raw input. Tasks are extracted from the message stream automatically (or inferred from explicit outcome reporting).
  45. - **Task complete or failed** β€” When a task is marked done or failed (e.g. by agent report or automatic detection), that outcome is the trigger for learning.
  46. - **Distillation** β€” An LLM pass infers from the conversation and execution trace what worked, what failed, and user preferences.
  47. - **Skill Agent** β€” Decides where to store (existing skill or new) and writes according to your `SKILL.md` schema.
  48. - **Update Skills** β€” Skills are updated. You define the structure in `SKILL.md`; the system does extraction, routing, and writing.
  49. ### Recall β€” How the agent uses skills on the next run
  50. ```mermaid
  51. flowchart LR
  52. E[Any Agent] --> F[list_skills/get_skill]
  53. F --> G[Appear in context]
  54. ```
  55. Give your agent **Skill Content Tools** (`get_skill`, `get_skill_file`). The agent decides what it needs, calls the tools, and gets the skill content. No embedding search β€” **progressive disclosure, agent in the loop**.
  56. # πŸͺœ Use It to Improve your Agent
  57. Claude Code:
  58. ```text
  59. Read https://acontext.io/SKILL.md and follow the instructions to install and configure Acontext for Claude Code
  60. ```
  61. OpenClaw:
  62. ```text
  63. Read https://acontext.io/SKILL.md and follow the instructions to install and configure Acontext for OpenClaw
  64. ```
  65. # πŸš€ Step-by-step Quickstart
  66. ### Connect to Acontext
  67. 1. Go to [Acontext.io](https://acontext.io), claim your free credits.
  68. 2. Go through a one-click onboarding to get your API Key (starts with `sk-ac`)
  69. <div align="center">
  70. <picture>
  71. <img alt="Dashboard" src="./assets/onboard.png" width="80%">
  72. </picture>
  73. </div>
  74. <details>
  75. <summary>πŸ’» Self-host Acontext</summary>
  76. We have an `acontext-cli` to help you do a quick proof-of-concept. Download it first in your terminal:
  77. ```bash
  78. curl -fsSL https://install.acontext.io | sh
  79. ```
  80. You should have [docker](https://www.docker.com/get-started/) installed and an OpenAI API Key to start an Acontext backend on your computer:
  81. ```bash
  82. mkdir acontext_server && cd acontext_server
  83. acontext server up
  84. ```
  85. > Make sure your LLM has the ability to [call tools](https://platform.openai.com/docs/guides/function-calling). By default, Acontext will use `gpt-4.1`.
  86. `acontext server up` will create/use `.env` and `config.yaml` for Acontext, and create a `db` folder to persist data.
  87. Once it's done, you can access the following endpoints:
  88. - Acontext API Base URL: http://localhost:8029/api/v1
  89. - Acontext Dashboard: http://localhost:3000/
  90. </details>
  91. ### Install SDKs
  92. We're maintaining Python [![pypi](https://img.shields.io/pypi/v/acontext.svg)](https://pypi.org/project/acontext/) and Typescript [![npm](https://img.shields.io/npm/v/@acontext/acontext.svg?logo=npm&logoColor=fff&style=flat&labelColor=2C2C2C&color=28CF8D)](https://www.npmjs.com/package/@acontext/acontext) SDKs. The snippets below are using Python.
  93. > Click the doc link to see TS SDK Quickstart.
  94. ```bash
  95. pip install acontext
  96. ```
  97. ### Initialize Client
  98. ```python
  99. import os
  100. from acontext import AcontextClient
  101. # For cloud:
  102. client = AcontextClient(
  103. api_key=os.getenv("ACONTEXT_API_KEY"),
  104. )
  105. # For self-hosted:
  106. client = AcontextClient(
  107. base_url="http://localhost:8029/api/v1",
  108. api_key="sk-ac-your-root-api-bearer-token",
  109. )
  110. ```
  111. ### Skill Memory in Action
  112. Create a learning space, attach a session, and let the agent learn β€” skills are written as Markdown files automatically.
  113. ```python
  114. from acontext import AcontextClient
  115. client = AcontextClient(api_key="sk-ac-...")
  116. # Create a learning space and attach a session
  117. space = client.learning_spaces.create()
  118. session = client.sessions.create()
  119. client.learning_spaces.learn(space.id, session_id=session.id)
  120. # Run your agent, store messages β€” when tasks complete, learning runs automatically
  121. client.sessions.store_message(session.id, blob={"role": "user", "content": "My name is Gus"})
  122. client.sessions.store_message(session.id, blob={"role": "assistant", "content": "Hi Gus! How can I help you today?"})
  123. # ... agent runs ...
  124. # List learned skills (Markdown files)
  125. client.learning_spaces.wait_for_learning(space.id, session_id=session.id)
  126. skills = client.learning_spaces.list_skills(space.id)
  127. # Download all skill files to a local directory
  128. for skill in skills:
  129. client.skills.download(skill_id=skill.id, path=f"./skills/{skill.name}")
  130. ```
  131. > `wait_for_learning` is a blocking helper for demo purposes. In production, task extraction and learning run in the background automatically β€” your agent never waits.
  132. ### More Features
  133. - **[Context Engineering](https://docs.acontext.io/engineering/editing)** β€” Compress context with summaries and edit strategies
  134. - **[Disk](https://docs.acontext.io/store/disk)** β€” Virtual, persistent filesystem for agents
  135. - **[Sandbox](https://docs.acontext.io/store/sandbox)** β€” Isolated code execution with bash, Python, and [mountable skills](https://docs.acontext.io/tool/bash_tools#mounting-skills-in-sandbox)
  136. - **[Agent Tools](https://docs.acontext.io/tool/whatis)** β€” Disk tools, sandbox tools, and skill tools for LLM function calling
  137. # 🧐 Use Acontext to Build Agents
  138. Download end-to-end scripts with `acontext`:
  139. **Python**
  140. ```bash
  141. acontext create my-proj --template-path "python/openai-basic"
  142. ```
  143. More examples on Python:
  144. - `python/openai-agent-basic`: openai agent sdk template
  145. - `python/openai-agent-artifacts`: agent can edit and download artifacts
  146. - `python/claude-agent-sdk`: claude agent sdk with `ClaudeAgentStorage`
  147. - `python/agno-basic`: agno framework template
  148. - `python/smolagents-basic`: smolagents (huggingface) template
  149. - `python/interactive-agent-skill`: interactive sandbox with mountable agent skills
  150. **Typescript**
  151. ```bash
  152. acontext create my-proj --template-path "typescript/openai-basic"
  153. ```
  154. More examples on Typescript:
  155. - `typescript/vercel-ai-basic`: agent in @vercel/ai-sdk
  156. - `typescript/claude-agent-sdk`: claude agent sdk with `ClaudeAgentStorage`
  157. - `typescript/interactive-agent-skill`: interactive sandbox with mountable agent skills
  158. > [!NOTE]
  159. >
  160. > Check our example repo for more templates: [Acontext-Examples](https://github.com/memodb-io/Acontext-Examples).
  161. >
  162. # πŸ” Documentation
  163. To learn more about skill memory and what Acontext can do, visit [our docs](https://docs.acontext.io/) or start with [What is Skill Memory?](https://docs.acontext.io/learn/quick)
  164. # ❀️ Stay Updated
  165. Star Acontext on GitHub to support us and receive instant notifications.
  166. ![click_star](./assets/star_acontext.gif)
  167. # πŸ—οΈ Architecture
  168. <details>
  169. <summary>click to open</summary>
  170. ```mermaid
  171. graph TB
  172. subgraph "Client Layer"
  173. PY["pip install acontext"]
  174. TS["npm i @acontext/acontext"]
  175. end
  176. subgraph "Acontext Backend"
  177. subgraph " "
  178. API["API<br/>localhost:8029"]
  179. CORE["Core"]
  180. API -->|FastAPI & MQ| CORE
  181. end
  182. subgraph " "
  183. Infrastructure["Infrastructures"]
  184. PG["PostgreSQL"]
  185. S3["S3"]
  186. REDIS["Redis"]
  187. MQ["RabbitMQ"]
  188. end
  189. end
  190. subgraph "Dashboard"
  191. UI["Web Dashboard<br/>localhost:3000"]
  192. end
  193. PY -->|RESTFUL API| API
  194. TS -->|RESTFUL API| API
  195. UI -->|RESTFUL API| API
  196. API --> Infrastructure
  197. CORE --> Infrastructure
  198. Infrastructure --> PG
  199. Infrastructure --> S3
  200. Infrastructure --> REDIS
  201. Infrastructure --> MQ
  202. style PY fill:#3776ab,stroke:#fff,stroke-width:2px,color:#fff
  203. style TS fill:#3178c6,stroke:#fff,stroke-width:2px,color:#fff
  204. style API fill:#00add8,stroke:#fff,stroke-width:2px,color:#fff
  205. style CORE fill:#ffd43b,stroke:#333,stroke-width:2px,color:#333
  206. style UI fill:#000,stroke:#fff,stroke-width:2px,color:#fff
  207. style PG fill:#336791,stroke:#fff,stroke-width:2px,color:#fff
  208. style S3 fill:#ff9900,stroke:#fff,stroke-width:2px,color:#fff
  209. style REDIS fill:#dc382d,stroke:#fff,stroke-width:2px,color:#fff
  210. style MQ fill:#ff6600,stroke:#fff,stroke-width:2px,color:#fff
  211. ```
  212. </details>
  213. # 🀝 Stay Together
  214. Join the community for support and discussions:
  215. - [Follow Acontext on X](https://x.com/alook_ai) 𝕏
  216. # 🌟 Contributing
  217. - Check our [roadmap.md](./ROADMAP.md) first.
  218. - Read [contributing.md](./CONTRIBUTING.md)
  219. # πŸ₯‡ Badges
  220. ![Made with Acontext](./assets/badge-made-with-acontext.svg) ![Made with Acontext (dark)](./assets/badge-made-with-acontext-dark.svg)
  221. ```md
  222. [![Made with Acontext](https://assets.memodb.io/Acontext/badge-made-with-acontext.svg)](https://acontext.io)
  223. [![Made with Acontext](https://assets.memodb.io/Acontext/badge-made-with-acontext-dark.svg)](https://acontext.io)
  224. ```
  225. # πŸ“‘ LICENSE
  226. This project is currently licensed under [Apache License 2.0](LICENSE).

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.