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

# Huntress

> How the agent reads each customer's Huntress escalations, incident reports, hosts and identities, and, after a technician approves, resolves escalations, acts on the SOC's remediation plan and isolates or releases hosts.

Connect your Huntress account and the agent works the tickets Huntress opens in your PSA, without a technician opening the Huntress portal.

When Huntress opens a ticket for an unexpected login, a noncompliant Microsoft 365 setting or a confirmed incident, the agent finds the escalation or incident report behind it, reads what Huntress saw, and writes it up in an internal note. It then proposes the next step: mark the login expected or unauthorized, approve the SOC's remediation plan, or isolate or release a host. Nothing changes in Huntress until a technician approves the plan.

## Prerequisites

* A Huntress account with your customers set up as **organizations**.
* A Huntress user for the key to act as. The key carries that user's role:
  * **Admin**, if the agent may make approved changes (resolve escalations, approve remediations, isolate hosts).
  * **Read-only**, if the agent should only read.

## How to connect

**1. Create a user API key**

Sign in to Huntress as an Admin. Open the menu at the top right and choose **API Credentials**. Under **User API Credentials**, click **Add**, pick the Huntress user the key acts as, and name the key (for example, Rallied).

**2. Copy the key and secret**

Copy the **API Key** and the **API Secret Key** right away. Huntress shows the secret only once.

**3. Connect Huntress**

In Rallied, go to **Settings → Integrations**, find the Huntress card and click **Connect**. Paste the **API Key** and **API Secret Key** and click **Connect**. The card shows **Connected** when it's ready.

Use a **user** API key. Rallied rejects the older account API key, because Huntress has deprecated it: it is read-only and does not return all data.

If you lost the secret, create a new key in Huntress and reconnect.

## Mapping organizations to clients

When you connect, Rallied syncs the organizations in your Huntress account. Review how each organization maps to a Rallied client on the **Companies** tab of the integration's detail page. Click **Sync organizations** to pick up customers you've added since.

On a client's ticket, the agent reads and changes only that client's organization. It fills in the organization ID automatically. A record that belongs to another customer, or to several customers at once, is answered as not found.

In the MSP admin [Chat](https://docs.rallied.ai/dashboard/chat.md), the agent can also list across every organization and read account-level data: all organizations, everyone with access to your Huntress account, invoices, and Managed SIEM log searches. End users can't run any Huntress tool.

## How the agent uses it

Huntress records the PSA ticket ID on what it opened, so the agent finds the record from the ticket itself.

**Unexpected login (ITDR).** For a ticket like "Unexpected Country" or "Unexpected VPN", the agent:

1. Finds the escalation and reads the user, country, VPN or IP, and time.
2. Reads the user's identity in Huntress (MFA, risk level) and any access rules already set for the customer.
3. Writes an internal note with what it found and proposes one resolution for approval:
   * **Expected**, for that user or for the whole customer. It can expire on a date, for example after a trip.
   * **Unauthorized**, for that user. It can also sign the user out and disable the account, only when the evidence points to a compromise.

**Incident report (EDR or identity).** The agent summarizes the SOC's findings and remediation plan in an internal note. Then it proposes, in one plan:

* approving the plan, which starts the steps Huntress runs itself, or rejecting it with a comment that goes back to the Huntress SOC;
* the manual steps for a technician, and marking each one done;
* releasing the host or identity once it is clean.

When the whole plan is approved, Huntress closes the incident report on its own and undoes its containment: an isolated host is released and a disabled identity is enabled again.

**Agent down.** For an RMM alert that the Huntress agent stopped on a computer, the agent looks the computer up in Huntress and reports whether it is still checking in, has stopped reporting, or is not enrolled in the customer's organization.

**Microsoft 365 settings (ISPM) and platform health.** These are fixed in the customer's tenant or devices, not in Huntress. The agent says which setting or devices are affected and what the recommended state is. It proposes resolving the escalation only after the fix is in place.

## What to know

* **Isolating a host cuts it off the network** until it is released. The agent proposes it only when the SOC or the evidence calls for it.
* **Comments on an incident report** are visible to everyone who can see the report in Huntress, which can include the customer's own Huntress users. Comments do not reach the SOC. Only rejecting the remediation plan does.
* **Changes that reach every customer** are only available in the admin chat. That covers creating a new organization, resolving an escalation for the whole account, and access rules for the whole account. On a client's ticket, a plan with one of these steps is refused before anyone approves it.
* **Invitations** to a customer's organization don't show in Huntress' API until the person accepts. Until then, the agent can't change or remove that person's access.
* **Permission errors** that name a Huntress role mean the key's user lacks that permission. Give the user a role with it, or leave the change to someone who has it.

## Tools

Reads default to **Auto**. Every change defaults to **Approval needed**. To change a default, open the **Policies** tab on **Settings → Integrations → Huntress**. See [Approval settings](https://docs.rallied.ai/dashboard/policies.md).

### Reads

| Tool | Default policy | Description |
| - | - | - |
| `huntress:huntress-list-escalations` | Auto | List a customer's escalations: unexpected logins, noncompliant settings, missing EDR |
| `huntress:huntress-get-escalation` | Auto | Get one escalation with the users, hosts or settings it is about |
| `huntress:huntress-list-incident-reports` | Auto | List a customer's incident reports (confirmed threats) |
| `huntress:huntress-get-incident-report` | Auto | Get one incident report with the SOC's findings |
| `huntress:huntress-list-incident-report-comments` | Auto | List the comments on one incident report |
| `huntress:huntress-list-remediations` | Auto | List the remediation steps the SOC proposed for one incident and their status |
| `huntress:huntress-list-agents` | Auto | List the hosts Huntress protects for a customer, with last check-in and isolation state |
| `huntress:huntress-get-agent` | Auto | Get one host: check-in, isolation, Defender and tamper protection |
| `huntress:huntress-list-identities` | Auto | List the user accounts Huntress ITDR watches, with risk and MFA state |
| `huntress:huntress-get-identity` | Auto | Get one user account Huntress ITDR watches |
| `huntress:huntress-list-signals` | Auto | List the behaviors Huntress investigated for a customer |
| `huntress:huntress-get-signal` | Auto | Get one behavior Huntress investigated |
| `huntress:huntress-list-platform-actions` | Auto | List configuration or health problems Huntress reports for a customer |
| `huntress:huntress-get-platform-action` | Auto | Get one configuration or health problem |
| `huntress:huntress-list-summary-reports` | Auto | List a customer's monthly, quarterly or yearly summary reports |
| `huntress:huntress-get-summary-report` | Auto | Get one summary report |
| `huntress:huntress-list-external-ports` | Auto | List the internet-facing ports External Recon found for a customer |
| `huntress:huntress-list-remote-access-connections` | Auto | List the remote access tools Huntress saw on a customer's computers |
| `huntress:huntress-list-unwanted-access-rules` | Auto | List which countries, VPNs and IPs are expected or unauthorized for a customer |
| `huntress:huntress-list-known-vpns` | Auto | List the VPN and proxy services Huntress recognizes |
| `huntress:huntress-list-siem-log-sources` | Auto | List the log sources feeding Managed SIEM for a customer |
| `huntress:huntress-list-misconfigured-log-sources` | Auto | List the log sources Managed SIEM reports as misconfigured |
| `huntress:huntress-get-organization` | Auto | Get the customer's Huntress organization with its counts and products |
| `huntress:huntress-list-organization-users` | Auto | List who has access to a customer's organization and with which role |
| `huntress:huntress-list-organizations` | Auto | List every organization in your Huntress account. Admin chat only |
| `huntress:huntress-get-account` | Auto | Get your Huntress account and the user the connection uses. Admin chat only |
| `huntress:huntress-list-account-users` | Auto | List everyone with access to your Huntress account. Admin chat only |
| `huntress:huntress-list-invoices` | Auto | List your Huntress invoices. Admin chat only |
| `huntress:huntress-query-siem-logs` | Auto | Search Managed SIEM logs with a query. Admin chat only |

### Changes

| Tool | Default policy | Description |
| - | - | - |
| `huntress:huntress-set-escalation-resolution` | Approval needed | Resolve an escalation, for example marking an unexpected login expected or unauthorized |
| `huntress:huntress-set-remediations-approved` | Approval needed | Approve the SOC's remediation plan, which starts the steps Huntress runs itself |
| `huntress:huntress-set-remediations-rejected` | Approval needed | Reject the remediation plan and send it back to the SOC with a comment |
| `huntress:huntress-set-remediation-completion` | Approval needed | Mark a manual remediation step as done or not done |
| `huntress:huntress-set-incident-report-resolved` | Approval needed | Mark an incident report as resolved |
| `huntress:huntress-add-incident-report-comment` | Approval needed | Add a comment to an incident report |
| `huntress:huntress-start-agent-isolation` | Approval needed | Cut a computer off the network (containment) |
| `huntress:huntress-stop-agent-isolation` | Approval needed | Give an isolated computer its network back |
| `huntress:huntress-update-agent` | Approval needed | Change a host's tags or tamper protection |
| `huntress:huntress-uninstall-agent` | Approval needed | Remove Huntress from a computer |
| `huntress:huntress-create-unwanted-access-rule` | Approval needed | Mark logins from a country, VPN or IP as expected or unauthorized for a customer |
| `huntress:huntress-update-unwanted-access-rule` | Approval needed | Change the dates or note of an access rule |
| `huntress:huntress-delete-unwanted-access-rule` | Approval needed | Remove an access rule |
| `huntress:huntress-create-organization` | Approval needed | Create a new customer organization. Admin chat only |
| `huntress:huntress-update-organization` | Approval needed | Rename a customer's organization or change who gets its reports |
| `huntress:huntress-delete-organization` | Approval needed | Remove a customer's organization (offboarding) |
| `huntress:huntress-invite-organization-user` | Approval needed | Give a person access to a customer's organization |
| `huntress:huntress-update-organization-user` | Approval needed | Change what a person can do in a customer's organization |
| `huntress:huntress-delete-organization-user` | Approval needed | Take away a person's access to a customer's organization |

Deleting an organization and uninstalling an agent can't be undone from Rallied.

## See also

* [Approval settings](https://docs.rallied.ai/dashboard/policies.md)
* [Chat](https://docs.rallied.ai/dashboard/chat.md)
* [Integrations](https://docs.rallied.ai/dashboard/integrations.md)
* [Integrations overview](https://docs.rallied.ai/integrations/overview.md)


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