> For the complete documentation index, see [llms.txt](https://docs.groundcover.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.groundcover.com/use-groundcover/connectors/claude-managed-agents.md).

# Claude Managed Agents

The Claude Managed Agents connector lets [Agent Mode](/use-groundcover/agent-mode.md) delegate coding tasks to a Claude agent configured in the Anthropic Console. Unlike connectors such as [Notion](/use-groundcover/connectors/notion.md) and [GitHub](/use-groundcover/connectors/github.md), groundcover communicates with this connector through the Claude Managed Agents API rather than an MCP server. The configured Claude agent can still use its own MCP servers and tools.

## Prerequisites

* Access to [Claude Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview)
* A Claude API key from the [Anthropic Console](https://platform.claude.com/)
* An [Agent ID](https://platform.claude.com/docs/en/managed-agents/agent-setup)
* An [Environment ID](https://platform.claude.com/docs/en/managed-agents/environments)
* Optional [vault IDs](https://platform.claude.com/docs/en/managed-agents/vaults) and a [memory store](https://platform.claude.com/docs/en/managed-agents/memory)

## Admin: Enable the Connector

1. Go to **Integrations → Connectors → Org Connectors**.
2. Select **Claude Managed Agents**.
3. Toggle the connector to **Enabled**.

No organization-level Claude credentials are required. Enabling the connector makes it available under **My Connections**.

## Connect a Claude Managed Agent

1. Go to **Integrations → Connectors → My Connections**.
2. Select **Claude Managed Agents**.
3. Enter the required values:
   * **API Key**: Your Claude API key.
   * **Agent ID**: The Claude agent that should run tasks.
   * **Environment ID**: The environment in which the agent should run.
4. Optionally add one or more **Vault IDs**.
5. Optionally expand **Memory Store** and configure a memory store.
6. Click **Connect**.

You can add multiple Claude Managed Agents connections. After the first connection, each additional connection requires a unique name so Agent Mode can distinguish them.

## Optional Configuration

{% hint style="warning" %}
Memory-store access initially defaults to **Read & write**. Explicitly select **Read only** for shared reference data and whenever the agent does not need to modify memory. Only keep **Read & write** when the agent deliberately needs to update the store: prompt injection from untrusted input can otherwise persist malicious instructions and affect later sessions.
{% endhint %}

| Setting             | Description                                                                                              |
| ------------------- | -------------------------------------------------------------------------------------------------------- |
| **Vault IDs**       | Secrets that the configured Claude agent can use while running tasks                                     |
| **Memory Store ID** | A memory store that persists context across agent runs                                                   |
| **Access**          | Whether the agent has **Read & write** (the initial default) or **Read only** access to the memory store |
| **Instructions**    | Guidance describing how the agent should use the memory store                                            |

## Manage a Connection

Open a Claude connection under **My Connections** to:

* Edit its Agent ID, Environment ID, vaults, or memory store.
* Rename it.
* Update the API key without recreating the connection.
* Click **Test connection** to send a short probe and confirm that Claude can reach the configured agent and environment.
* Configure tool restrictions to run actions automatically, require approval, or deny them.
* Add another Claude connection or disconnect the current one.

## Troubleshooting

| Issue                                                     | Resolution                                                                                                  |
| --------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------- |
| Claude Managed Agents is unavailable under My Connections | Ask a workspace admin to enable it under **Org Connectors**                                                 |
| Test connection fails                                     | Confirm the API key is active and the Agent ID and Environment ID belong to resources the key can access    |
| A task cannot access a secret                             | Confirm the correct vault ID is attached and that the configured agent has permission to use it             |
| Memory is not retained                                    | Confirm the Memory Store ID and access level, and check that the store is available to the configured agent |


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.groundcover.com/use-groundcover/connectors/claude-managed-agents.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.