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

# Jira

> Connect a Jira Cloud site so the agent can open, comment on, and transition issues.

Nuphos connects to Jira Cloud through **Atlassian OAuth** — you authorize once
in the browser and there's no API token or site URL to manage. The agent can
then work with issues on your site: file a ticket for an incident, comment
progress, and transition it to Done when the fix lands.

## Connect

<Steps>
  <Step title="Start the authorization">
    In **Connectors → Connect → Jira**, click **Authorize Jira**. Your
    browser opens Atlassian's consent screen.
  </Step>

  <Step title="Choose one site and approve">
    Grant access to a single Jira site. Nuphos asks for `read:jira-work` and
    `write:jira-work` to read, create, and update issues and comments, and
    `read:jira-user` to resolve assignees. Note that `read:jira-user` allows
    reading the user profiles on your site — display names, email addresses
    (where visible), and avatars.
  </Step>

  <Step title="Return to Nuphos">
    Atlassian redirects back to the desktop app and the connection completes
    automatically.
  </Step>
</Steps>

Tokens are stored encrypted and refreshed automatically. Re-authorizing the
same site updates the existing connection instead of creating a duplicate.

<Info>
  Jira **Cloud** only — Data Center / self-hosted Jira isn't supported. Each
  connection covers one site; if your grant spans several sites, Nuphos asks
  you to re-authorize with a single site selected.
</Info>

## In the Nuphos app

There's no separate Jira browsing surface — the connector's value shows up
inside [agent sessions](/agent/sessions), where issues the agent files or
updates are linked in the conversation. The binding itself lives in
**Connectors**, where an Administrator can manage its allow-list or remove it.

## What the agent can do for you

The agent can discover projects and issue types, create issues, comment, search
with JQL, and move issues through your workflow transitions. The pattern that
makes this shine is pairing it with your infrastructure connectors:

* *"Investigate the elevated error rate — and open a Jira ticket with what you
  find."* — the agent diagnoses via your cloud and observability connectors,
  then files the issue with the evidence.
* As it works a fix, it **comments progress** on the ticket, and when the fix
  lands it transitions the issue through your workflow to Done.
* *"What's assigned to me about the payment service?"* — it searches with JQL
  and summarizes.

## Permissions & authorization

Two boundaries govern what happens through this connector:

* **Hard boundary — who can use it.** Only team **Administrators** can connect
  or remove the binding, and they can restrict it to specific members with a
  [member allow-list](/security/access-control). Within a session the agent only
  holds credentials you explicitly selected, re-checked on every request.
* **Soft boundary — what the agent may do with it.** Reads are free-form;
  creating, commenting, and transitioning go through
  [plan approvals](/security/approvals) unless you gave an explicit instruction
  (e.g. "file a ticket for this"), with
  [Auto-authorization](/security/auto-mode) governing the individual actions.

The outer limit is the OAuth scopes above — issue read/write and user lookup on
one site, and nothing beyond it.

## Good to know

* One connection covers **one Jira site** — connect again for additional sites.
* The agent works within your existing **workflow transitions**; it can't skip
  states your workflow doesn't allow.