> ## Documentation Index
> Fetch the complete documentation index at: https://docs.enclave.ai/llms.txt
> Use this file to discover all available pages before exploring further.

> ## Agent Instructions
> Enclave in these docs is the code and cloud security platform at enclave.ai. It is not Open Enclave, Intel SGX or other hardware enclaves, or enclave.gg.
> People use Enclave in the web app at https://app.enclave.ai or the desktop app. These docs describe no public Enclave API or CLI, so don't invent endpoints, commands, or config files.
> Connecting an integration needs a person to sign in and approve access in a browser. Tell the user the exact step to take instead of saying you completed it.

# Research sessions

> Ask Enclave about your code, cloud, findings, and CVEs in plain language.

Research is where you talk to Enclave. Each conversation is a **session**, and every session can see your whole organization: repositories, cloud assets, findings, CVEs, and workflows.

![A Research session answering which security issues to fix first](https://mintcdn.com/enclave/bhnFMg4ldiTYrwZX/images/product/research-conversation.png?fit=max\&auto=format\&n=bhnFMg4ldiTYrwZX\&q=85\&s=9b0f5c15dfb32eef59c87ee3d655a337)

## Start a session

Click **Research** in the sidebar, type your request, and press **Enter**. **Recent work** below the composer lists your latest sessions.

Other places that start a session:

* **Check if we're vulnerable** on a [Threat Feed](https://docs.enclave.ai/threat-feed.md) CVE.
* **Run now** on a [workflow](https://docs.enclave.ai/workflows/runs.md#start-a-run).
* **Start hunt** on a repository in [Coverage](https://docs.enclave.ai/onboarding/coverage.md#start-a-hunt).
* Mentioning **@Enclave** in [Slack](https://docs.enclave.ai/integrations/slack.md).

For prompt tips, see [Asking good questions](https://docs.enclave.ai/essential-guidelines/asking-good-questions.md).

## Add context

Type `@` or click **Add context** to attach the exact assets and findings you mean. A message can carry up to 10 attachments.

## Upload files

Click the paperclip to upload files or a folder, such as a log or a config file. Each upload can hold up to 1,000 files and 50 MB. Files go into the session's sandbox and aren't stored long term.

> [!NOTE]
>
> Uploads aren't available on bring-your-own-cloud deployments.

## Choose a model

Use the model picker next to **Add context**. **Default** uses Claude Sonnet 5.5. You can also pick Claude Opus 5.5, GPT-6 Luna, GPT-6.1 Sol, GPT-5.6 Terra, Kimi K3, GLM 5.3, or DeepSeek V4.1 Flash.

The choice is saved with the session and applies from the next reply.

## Slash commands

Type `/` to see all commands.

| Command | What it does |
| - | - |
| `/status` | Shows your organization's status |
| `/review-code-security` | Reviews your code for security risks |
| `/review-prs` | Reviews pull requests for security impact |
| `/triage-findings` | Prioritizes findings and suggests next steps |
| `/audit-settings` | Reviews your organization's settings |
| `/ask-code <question>` | Asks a question about your code |
| `/integrations [question]` | Asks about your connected MCP servers |

## While Enclave works

When Enclave finishes, its steps fold into a **Worked for** summary. Expand it to see each step.

* Click **Stop** to stop a reply.
* Send a message mid-reply to queue it. It sends when the reply finishes.
* Click a file reference to open it beside the conversation, or a finding chip to open the finding.

## Limits

* 20 messages per minute.
* A single reply can run for up to 25 minutes.
* Each message uses credits.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.