Skip to content
Back to search

Adrian

Open-source runtime AI agent security tool - monitors and controls AI agents, catching malicious tool use, prompt injection, and policy drift in real time, before the agent acts.

Safety scan

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

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/secureagentics--adrian/badge.svg)](/listings/secureagentics--adrian)

HTML

<a href="/listings/secureagentics--adrian"><img src="/api/listings/secureagentics--adrian/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

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

> **Want the stupidly simple, 60-second hands-off install?** Feed your coding agent (Claude, Codex, Cursor, etc.) this file: [GET_STARTED_AI_GUIDE.md](https://github.com/secureagentics/Adrian/blob/main/GET_STARTED_AI_GUIDE.md). It will walk you through the installation process ([video guide here](https://youtu.be/7vYjeGxY8to)). Always review instructions manually.

The next fastest way to try Adrian is the managed dashboard at [app.adrian.secureagentics.ai](https://app.adrian.secureagentics.ai). Sign-up takes a minute and there is nothing to install beyond the SDK. To run Adrian on your own infrastructure instead, jump to [Self-hosting](#self-hosting) below.

1. Sign up at [app.adrian.secureagentics.ai](https://app.adrian.secureagentics.ai) and generate an API key.

2. Configure Adrian for your agent and your preferences (remit of your agent, audit vs block mode, alerting channels, accepted behaviours vs known-risks).

3. Install the SDK:

   pip install adrian-sdk

4. Install LangChain and the provider for your agent's model (the SDK auto-instruments LangChain / LangGraph; pick whichever provider matches your model):

   pip install langchain langchain-openai   # or langchain-anthropic, etc.
   # or, in a uv project:  uv add langchain langchain-openai

<sup>`langchain` pulls `langgraph` in, so this covers both `create_agent` and `create_react_agent`. Last verified 2026-06-24 with `langchain==1.3.9`, `langgraph==1.2.5`, `langchain-core==1.4.7`, `langchain-openai==1.3.2`. Supported: `langchain`/`langgraph`/`langchain-openai` `>=1.0,<2.0`, `langchain-core` `>=1.2.19,<2.0`.</sup>

5. Wrap your LangChain agent. Two lines of Adrian (`init` + `shutdown`) bracket your normal LangChain / LangGraph code:

   import asyncio
   import adrian
   from langchain_openai import ChatOpenAI

   async def main():
       adrian.init(api_key="adr_live_...")
       llm = ChatOpenAI(model="gpt-4o")
       response = await llm.ainvoke(
           "Find the most underpriced recent IPOs and build an investment strategy",
       )
       print(response.content)
       adrian.shutdown()

   asyncio.run(main())

Full runnable version (with env-var checks) at [`examples/python/quickstart.py`](examples/python/quickstart.py). More complex examples using agents are in [`examples/python/`](examples/python/).

6. Run your agent. Events appear in the dashboard within seconds, classified by severity.

Full guide: [Quickstart](https://docs.adrian.secureagentics.ai/quickstart).

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 · Manifest · 16 lines

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

  1. {
  2. "name": "adrian",
  3. "owner": {
  4. "name": "Secure Agentics",
  5. "email": "support@secureagentics.ai"
  6. },
  7. "description": "Runtime AI-security tooling for Claude Code and agentic AI, from the Adrian OSS project.",
  8. "plugins": [
  9. {
  10. "name": "adrian-cc",
  11. "source": "./integrations/claude-code",
  12. "description": "Runtime security for Claude Code: hooks-based observability and blocking, backed by the self-hosted Adrian engine.",
  13. "strict": true
  14. }
  15. ]
  16. }
README.md · README · 231 lines

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

  1. <p align="center">
  2. <picture>
  3. <source media="(prefers-color-scheme: dark)" srcset="assets/adrian-logo-dark.png">
  4. <img src="assets/adrian-logo-light.png" alt="Adrian by Secure Agentics" width="360">
  5. </picture>
  6. </p>
  7. <h3 align="center">Agent attacks slip past static analysis and network monitoring.</h3>
  8. <p align="center"><b>Adrian catches them at runtime, by watching what the agent actually does (its actions and its reasoning) and stepping in before it acts. Open source, free forever.</b></p>
  9. <p align="center"><b>+35% detection accuracy · 4x more nuanced attacks caught</b> vs behaviour-only monitoring (<a href="https://arxiv.org/pdf/2503.11926">OpenAI &amp; DeepMind research</a>)</p>
  10. <p align="center">
  11. <a href="https://github.com/secureagentics/Adrian/stargazers"><img src="https://img.shields.io/github/stars/secureagentics/Adrian?style=social" alt="Stars" /></a>
  12. <a href="LICENSE"><img src="https://img.shields.io/badge/licence-Apache--2.0-blue.svg" alt="Licence" /></a>
  13. <a href="https://app.adrian.secureagentics.ai/"><img src="https://img.shields.io/badge/Dashboard-Sign%20Up-22C55E" alt="Dashboard" /></a>
  14. <a href="https://pypi.org/project/adrian-sdk/"><img src="https://img.shields.io/pypi/v/adrian-sdk.svg" alt="PyPI" /></a>
  15. <a href="https://discord.gg/Vq2VyYrw8Z"><img src="https://img.shields.io/badge/Discord-Join-5865F2?logo=discord&logoColor=white" alt="Discord" /></a>
  16. <a href="https://www.linkedin.com/company/secure-agentics"><img src="https://img.shields.io/badge/LinkedIn-Follow-0A66C2?logo=linkedin&logoColor=white" alt="LinkedIn" /></a>
  17. <a href="https://www.producthunt.com/products/adrian?embed=true&utm_source=badge-featured&utm_medium=badge&utm_campaign=badge-adrian" target="_blank" rel="noopener noreferrer"><img src="https://img.shields.io/badge/Product%20Hunt-Featured-DA552F?logo=producthunt&logoColor=white" alt="Product Hunt" /></a>
  18. </p>
  19. <p align="center"><i>⭐ If you think agents need a runtime security layer, star the repo. It is how new people find Adrian, and how we know to keep building it in the open.</i></p>
  20. ---
  21. Adrian is an open-source, [AARM-aligned](https://aarm.dev) runtime security monitoring and control engine for AI agents. It analyses both agent activity logs (tool calls, actions, outputs) and reasoning traces to detect malicious, misaligned, or out-of-remit behaviour, and optionally intervene in-flight. SDKs are available for Python ([LangChain](sdk/python/README.md), [Anthropic](sdk/python/ANTHROPIC.md)) and TypeScript ([sdk/typescript/README.md](sdk/typescript/README.md)), plus a native [Claude Code plugin](integrations/claude-code/README.md) that secures every tool call from your terminal.
  22. > **🆕 Claude Code plugin - now live.** Drop Adrian into Claude Code and every tool call is classified in real time, with risky actions blocked or held for your approval right in the terminal. No code changes: install with `/plugin marketplace add secureagentics/Adrian` then `/adrian-init`. Full guide: **[integrations/claude-code/README.md](integrations/claude-code/README.md)**.
  23. <p align="center">
  24. <a href="https://docs.adrian.secureagentics.ai">Documentation</a> &nbsp;•&nbsp;
  25. <a href="https://app.adrian.secureagentics.ai">Dashboard</a> &nbsp;•&nbsp;
  26. <a href="https://discord.gg/Vq2VyYrw8Z">Discord</a> &nbsp;•&nbsp;
  27. <a href="https://www.linkedin.com/company/secure-agentics">LinkedIn</a>
  28. </p>
  29. > **▶️ See it in action:** Adrian catches an agent going out-of-remit in real time and steps in before the action lands.
  30. https://github.com/user-attachments/assets/ba50e6e4-fe3e-47b2-aa69-2902e1ef2924
  31. <sup>New to Adrian? Check out the [Launch Video](https://www.youtube.com/watch?v=NkEISlRhyFs).</sup>
  32. ## Why Adrian is different
  33. Most agent monitoring stops at activity logs: APIs, MCP, DB interactions, tool calls, etc. Adrian enhances this by also analysing the agent's reasoning: understanding _why_ it took an action, under what context, and what it is planning on doing next. Combining behaviour and reasoning analysis like this is exactly what the [OpenAI and DeepMind research](https://arxiv.org/pdf/2503.11926) found catches far more, and Adrian is the first tool to put it into a deployable security control, free forever.
  34. **What it catches:**
  35. - Prompt injection and jailbreaks, direct and indirect
  36. - Tool poisoning and unsafe or off-policy tool calls
  37. - Data exfiltration and secret / credential leakage
  38. - Privilege escalation and out-of-remit actions
  39. Classifiers trained on prompt-injection datasets only catch what they have seen before. Adrian works differently: it holds a working understanding of what your agent is meant to do and judges each new action against that, correlated across the whole session. So when your e-commerce agent starts resetting user passwords, something no training set would flag, Adrian catches it.
  40. ## Quickstart
  41. > **Want the stupidly simple, 60-second hands-off install?** Feed your coding agent (Claude, Codex, Cursor, etc.) this file: [GET_STARTED_AI_GUIDE.md](https://github.com/secureagentics/Adrian/blob/main/GET_STARTED_AI_GUIDE.md). It will walk you through the installation process ([video guide here](https://youtu.be/7vYjeGxY8to)). Always review instructions manually.
  42. The next fastest way to try Adrian is the managed dashboard at [app.adrian.secureagentics.ai](https://app.adrian.secureagentics.ai). Sign-up takes a minute and there is nothing to install beyond the SDK. To run Adrian on your own infrastructure instead, jump to [Self-hosting](#self-hosting) below.
  43. 1. Sign up at [app.adrian.secureagentics.ai](https://app.adrian.secureagentics.ai) and generate an API key.
  44. 2. Configure Adrian for your agent and your preferences (remit of your agent, audit vs block mode, alerting channels, accepted behaviours vs known-risks).
  45. 3. Install the SDK:
  46. ```sh
  47. pip install adrian-sdk
  48. ```
  49. 4. Install LangChain and the provider for your agent's model (the SDK auto-instruments LangChain / LangGraph; pick whichever provider matches your model):
  50. ```sh
  51. pip install langchain langchain-openai # or langchain-anthropic, etc.
  52. # or, in a uv project: uv add langchain langchain-openai
  53. ```
  54. <sup>`langchain` pulls `langgraph` in, so this covers both `create_agent` and `create_react_agent`. Last verified 2026-06-24 with `langchain==1.3.9`, `langgraph==1.2.5`, `langchain-core==1.4.7`, `langchain-openai==1.3.2`. Supported: `langchain`/`langgraph`/`langchain-openai` `>=1.0,<2.0`, `langchain-core` `>=1.2.19,<2.0`.</sup>
  55. 5. Wrap your LangChain agent. Two lines of Adrian (`init` + `shutdown`) bracket your normal LangChain / LangGraph code:
  56. ```python
  57. import asyncio
  58. import adrian
  59. from langchain_openai import ChatOpenAI
  60. async def main():
  61. adrian.init(api_key="adr_live_...")
  62. llm = ChatOpenAI(model="gpt-4o")
  63. response = await llm.ainvoke(
  64. "Find the most underpriced recent IPOs and build an investment strategy",
  65. )
  66. print(response.content)
  67. adrian.shutdown()
  68. asyncio.run(main())
  69. ```
  70. Full runnable version (with env-var checks) at [`examples/python/quickstart.py`](examples/python/quickstart.py). More complex examples using agents are in [`examples/python/`](examples/python/).
  71. 6. Run your agent. Events appear in the dashboard within seconds, classified by severity.
  72. Full guide: [Quickstart](https://docs.adrian.secureagentics.ai/quickstart).
  73. ## Self-hosting
  74. Adrian supports entirely offline, data sovereign deployments using just a handful of docker commands. This repository ships everything needed to run the entire Adrian stack on a single host: the Go backend (WebSocket + dashboard API + AI engine), the Next.js dashboard, the Python SDK, and a Llama.cpp container that serves a local Gemma model. No managed cloud, no telemetry leaving the box.
  75. > **Hardware support:** Tested on NVIDIA GPUs with Gemma 4 (E2B / E4B) which is the model the bootstrap picker downloads by default. CPU-only is technically possible but will be slow on real workloads with those sized models.
  76. ### Prerequisites
  77. - A host with Docker + Docker Compose v2.
  78. - An **NVIDIA GPU** with recent CUDA driver and the [NVIDIA Container Toolkit](https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html) installed (for the bundled Llama.cpp classifier). ~10 GB free disk for the model.
  79. ### Bring-up
  80. 1. **Clone:**
  81. ```
  82. git clone https://github.com/secureagentics/Adrian
  83. cd Adrian
  84. ```
  85. 2. **Run bootstrap.** Creates `data/adrian.db`, applies migrations, generates a random admin password, and writes `.env`. With no `--gguf` flag, the bootstrap interactively offers to download the recommended on-device classifier (Gemma 4 E4B, ~5 GB, or E2B ~3 GB) into `./models/`.
  86. ```sh
  87. # Default: interactive picker downloads Gemma 4 E4B / E2B
  88. docker compose --profile setup run --rm setup bootstrap
  89. # Already have a GGUF under ./models/? Pass it by name
  90. docker compose --profile setup run --rm setup bootstrap \
  91. --gguf my-model.gguf
  92. ```
  93. 3. **Start the stack.**
  94. ```sh
  95. docker compose --profile llm up -d
  96. ```
  97. 4. **Open the dashboard.** Browse to `http://localhost:3000`. Sign in with `admin@localhost` plus the password the bootstrap printed; you'll be prompted to set a new one. Create an SDK API key and configure Adrian to monitor your specific agent from **Settings → Agents → New key**.
  98. 5. **Wrap your agent.** The SDK lives in-tree under `sdk/`. Install it into a fresh `.venv` via the bundled Make target (uses [uv](https://docs.astral.sh/uv/)):
  99. ```sh
  100. make sdk-install
  101. source .venv/bin/activate
  102. ```
  103. Install LangChain and the provider for your agent's model into the same venv:
  104. ```sh
  105. uv pip install "langchain>=1.0,<2.0" "langchain-openai>=1.0,<2.0" # swap langchain-openai for your model's provider
  106. ```
  107. <sup>`langchain` pulls `langgraph` in, so this covers both `create_agent` and `create_react_agent`. Last verified 2026-06-24 with `langchain==1.3.9`, `langgraph==1.2.5`, `langchain-core==1.4.7`, `langchain-openai==1.3.2`.</sup>
  108. Use the same `adrian.init` snippet as in the [Quickstart](#quickstart) above. The SDK defaults to `ws://localhost:8080/ws`, so a self-hosted setup needs nothing more than the API key - drop the `ws_url=` line.
  109. To [reset the admin password](https://docs.adrian.secureagentics.ai/reference/backend#reset-the-admin-password), [change the model](https://docs.adrian.secureagentics.ai/reference/backend#switch-the-local-gguf) and much more check out the dedicated [Docs site](https://docs.adrian.secureagentics.ai/).
  110. ## Architecture
  111. ```mermaid
  112. flowchart TD
  113. Agent[Agent runtime] --> SDK[Adrian SDK]
  114. SDK --> Backend[Adrian backend]
  115. Backend --> Classifier[Classifier model]
  116. Classifier --> Verdict{Verdict}
  117. Verdict --> Control[Control plane]
  118. Verdict -.->|"Alert /<br>Human Review /<br>Block"| Agent
  119. ```
  120. ## Integrations
  121. <table>
  122. <thead>
  123. <tr><th></th><th>Supported</th><th>On roadmap</th></tr>
  124. </thead>
  125. <tbody>
  126. <tr>
  127. <th align="left">Frameworks</th>
  128. <td>
  129. <a href="https://www.langchain.com/"><img height="32" src="https://cdn.simpleicons.org/langchain/1FA383" alt="LangChain"></a>&nbsp;&nbsp;
  130. <a href="https://platform.openai.com/docs/agents"><picture><source media="(prefers-color-scheme: dark)" srcset="assets/logos/openai-dark.svg"><img height="32" src="assets/logos/openai-light.svg" alt="OpenAI Agents SDK"></picture></a>&nbsp;&nbsp;
  131. <a href="https://claude.com/claude-code"><img height="32" src="https://cdn.simpleicons.org/claude/D97757" alt="Claude Code"></a>&nbsp;&nbsp;
  132. <a href="https://docs.anthropic.com/"><img height="32" src="https://cdn.simpleicons.org/anthropic/D97757" alt="Anthropic SDK"></a>
  133. </td>
  134. <td>
  135. <a href="https://www.crewai.com/"><img height="32" src="https://cdn.simpleicons.org/crewai/FF5A50" alt="CrewAI"></a>&nbsp;&nbsp;
  136. <a href="https://github.com/openclaw/openclaw"><img height="32" src="https://raw.githubusercontent.com/openclaw/openclaw/main/docs/assets/pixel-lobster.svg" alt="OpenClaw"></a>
  137. </td>
  138. </tr>
  139. <tr>
  140. <th align="left">Alerting</th>
  141. <td>
  142. <a href="https://discord.com/"><img height="32" src="https://cdn.simpleicons.org/discord/5865F2" alt="Discord"></a>&nbsp;&nbsp;
  143. <a href="https://slack.com/"><img height="32" src="assets/logos/slack.svg" alt="Slack"></a>
  144. </td>
  145. <td>
  146. <a href="https://www.whatsapp.com/"><img height="32" src="https://cdn.simpleicons.org/whatsapp/25D366" alt="WhatsApp"></a>&nbsp;&nbsp;
  147. <a href="https://www.microsoft.com/microsoft-teams/group-chat-software"><img height="32" src="assets/logos/teams.svg" alt="Microsoft Teams"></a>&nbsp;&nbsp;
  148. <a href="https://www.pagerduty.com/"><img height="32" src="https://cdn.simpleicons.org/pagerduty/06AC38" alt="PagerDuty"></a>
  149. </td>
  150. </tr>
  151. </tbody>
  152. </table>
  153. Full list: [Integrations](https://docs.adrian.secureagentics.ai/integrations). **New:** the [Claude Code plugin](integrations/claude-code/README.md) is live. ⭐ the repo or [join Discord](https://discord.gg/Vq2VyYrw8Z).
  154. ## Contributing
  155. ⭐ Star the repo if Adrian is useful. Then see [CONTRIBUTING.md](CONTRIBUTING.md) for the full guide. In short: sign the [CLA](CLA.md), branch off `main`, follow the PR template, and use British English / no em-dashes in prose.
  156. See [CONTRIBUTORS.md](CONTRIBUTORS.md) for the list of people who have shaped Adrian, and how to add yourself.
  157. ## Licence
  158. Adrian is released under the [Apache 2.0 licence](LICENSE). New source files should carry the SPDX header from [LICENSE_HEADER.txt](LICENSE_HEADER.txt).
  159. ## Community
  160. - [Discord](https://discord.gg/Vq2VyYrw8Z) for chat with the team and other Adrian users
  161. - [LinkedIn](https://www.linkedin.com/company/secure-agentics) for product updates
  162. ## Featured on
  163. - [Product Hunt](https://www.producthunt.com/products/adrian?embed=true&utm_source=badge-featured&utm_medium=badge&utm_campaign=badge-adrian) <a href="https://www.producthunt.com/products/adrian?embed=true&utm_source=badge-featured&utm_medium=badge&utm_campaign=badge-adrian" target="_blank" rel="noopener noreferrer"><img src="https://cdn.simpleicons.org/producthunt/DA552F" alt="Featured on Product Hunt" height="25"></a>
  164. - [There's An AI For That](https://theresanaiforthat.com/ai/adrian/?ref=social-icon&v=10763736) <a href="https://theresanaiforthat.com/ai/adrian/?ref=social-icon&v=10763736" target="_blank" rel="nofollow"><img src="https://media.theresanaiforthat.com/social/icon_full.svg" alt="Featured on TAAFT" height="25"></a>

Ratings and comments

Sign in to rate this skill.

Sign in

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