# Anthropic's Claude Enterprise {% hint style="info" %} **Early Access** — This integration is available in early access. Contact your Snyk account team to enable it. {% endhint %} Integrate Snyk Evo with Anthropic's Claude Enterprise to ingest a snapshot of your Claude Enterprise organization. This integration surfaces AI assets, such as Claude models in use and organization-approved MCP servers, directly in your Evo inventory. ## Prerequisites * Snyk Evo with the Anthropic Claude Enterprise integration enabled. Contact your Snyk account team to request access. * A Claude Enterprise account. * Primary Owner role in Claude Enterprise. Only the Primary Owner can create a Compliance Access Key. ## Create a Compliance Access Key in Claude Enterprise 1. Log in to [claude.ai](https://claude.ai) as the Primary Owner of your organization. 2. Navigate to **Settings > Data and Privacy > Compliance access keys**. 3. Click **Create key**. 4. Enter a name for the key, for example, *Snyk Evo*. 5. Select the following three scopes: * `read:compliance_activities` * `read:compliance_user_data` * `read:compliance_org_data` 6. Click **Create**. 7. Copy the key and store it in a secure location. {% hint style="warning" %} You cannot change scopes after you create a key. If you select the wrong scopes, delete the key and create a new one. Do not select `delete:compliance_user_data`. Evo does not require this scope. {% endhint %} {% hint style="warning" %} Evo does not store your Compliance Access Key between sessions. Paste the key each time you run an ingestion. If you lose the key, return to Claude Enterprise and create a new one. {% endhint %} ## Ingest a snapshot in Evo 1. In Evo, navigate to **AI platform integrations**. 2. On the **Ingest Anthropic compliance snapshot** page, paste your Compliance Access Key into the **Compliance key** field. 3. Click **Ingest snapshot**. Evo calls Claude's Compliance API and ingests the models and MCP servers of your Claude Enterprise organization as native Evo assets. When the ingestion completes, a **Snapshot ingested** confirmation appears, and the page displays **Snapshot last ingested: Just now**. {% hint style="info" %} Ingestion is synchronous. For large workspaces, this process takes several seconds. {% endhint %} ## Refresh your data Evo does not automatically sync with Claude Enterprise. To update your inventory with the latest data, navigate to the **AI platform integrations** page, click **Ingest new snapshot**, and enter your Compliance Access Key. ## View your inventory After a successful ingestion, the **AI platform integrations** page displays two tabs: Models and MCP servers. ### Models The Models tab lists every Claude model detected in use across your organization. | Column | Description | | ------------- | -------------------------------------------------------------------- | | Name | Claude model identifier | | Added | When the model first appeared in your organization's data | | Projects | Number of Claude Projects associated with this model | | Organizations | Number of Claude Enterprise organizations where this model is active | Select a model to open its detail page. The detail page shows usage statistics (Organizations, Projects, Used in chats, Users) and three tabs: * Usage: Lists the organizations using this model, with the date added and the creator who first associated it. * Risk profile: Evo's AI Risk DB scores the model across vulnerability categories on a 0–1000 index (lower is better): | Category | Description | | ------------------------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------- | | Bias & discrimination | Measures fairness and discriminatory behavior across demographics. | | Insecure code generation | Measures how often the model generates code containing vulnerabilities such as SQL injection, XSS, OS injection, path traversal, or improper input validation. | | Sensitive data exposure | Tests whether the model memorizes and reproduces personally identifiable information (PII) from its training data. | | Attack reconnaissance | Measures susceptibility to assisting with information gathering for attacks. | | Safety guardrail bypass | Tests resistance to attempts to circumvent safety controls. | * Capabilities: Displays model metadata including vendor, model family, model variant, tasks, modalities, license type, cost tier, and a link to the Anthropic model documentation. ### MCP servers The MCP servers tab lists every MCP server added to your Claude Enterprise organization. | Column | Description | | ------------- | --------------------------------------------------------------------- | | Name | Name of the MCP server (for example, *PagerDuty*) | | Added | Date the server was added to the Claude Enterprise organization | | Projects | Number of Claude Projects associated with this server | | Organizations | Number of Claude Enterprise organizations where this server is active | Select an MCP server to open its detail page. The detail page shows usage statistics (Added, Organizations, Private projects, Public projects, Users) and two tabs: * Usages: Lists the organizations using this MCP server, with the date added and the creator's email. * Tools & permissions: Displays tool-level permission restrictions set by your Claude Enterprise administrators. The Tools table shows: | Column | Description | | --------------- | --------------------------------------------- | | Name | Tool name (for example, *delete\_task*) | | Date | When the restriction was set | | Set by | The administrator who applied the restriction | | Max permissions | Permission level for this tool | If no tool-level policies are configured in Claude Enterprise, this table shows **No data**. To set tool permissions, update your MCP server configuration in Claude Enterprise and click **Ingest new snapshot** again. ## Troubleshooting | Issue | Likely cause | Resolution | | ----------------------------------------- | --------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------- | | **Ingest snapshot** button is disabled | **Compliance key** field is empty | Paste your Compliance Access Key into the **Compliance key** field. | | Ingestion succeeds but Models tab shows 0 | No chat activity in your Claude Enterprise organization | Generate activity in Claude Enterprise, then click **Ingest new snapshot**. | | MCP servers tab shows 0 | No MCP servers are added to your Claude Enterprise organization | Add MCP servers in Claude Enterprise administrator settings, then click **Ingest new snapshot**. | | Ingestion fails | Key is invalid or revoked | Check the key status in Claude Enterprise under **Settings > Data and Privacy > Compliance access keys** and create a new key if needed. | | Ingestion fails with a scope error | Key lacks required scopes | Delete the key in Claude Enterprise and create a new one with all three required scopes. | --- # Agent Instructions: 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: ``` GET https://docs.snyk.io/integrations/cloud-ai-platforms/anthropic-claude-enterprise.md?ask= ``` The question should be specific, self-contained, and written in natural language. 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.