Jira, GitHub & Linear integrations

Task sources pull issues from Jira, GitHub, Linear or any HTTP endpoint into a project, so your team books time straight onto the tickets it already works with.

How task sources work

A task source is attached to a project and is strictly read-only: the server fetches the issue list on demand when someone opens the task picker, and nothing is ever written back to Jira, GitHub or Linear.

Picking a remote issue creates a matching task in the project the first time — named after the ticket and linked back to the original — and reuses it afterwards. That task creation emits a task.created webhook like any other:

#42 - Fix login redirect      ← GitHub issue 42
PROJ-7 - Fix login redirect   ← Jira issue PROJ-7
ENG-123 - Fix login redirect  ← Linear issue ENG-123

Connecting GitHub

Open Projects, expand a project card and choose Add source under Task sources, type GitHub. You need:

• An HTTPS repository URL, for example https://github.com/octocat/hello-world.
• A personal access token that can read the repository's issues.

The source lists the repository's issues; pull requests are filtered out.

Connecting Jira Cloud

Choose the Jira Cloud type. You need:

• Your site — the subdomain (acme) or the full URL (https://acme.atlassian.net).
• The email address of the Atlassian account.
• An Atlassian API token, created at id.atlassian.com.
• Optionally a project key to limit the source to one Jira project.

Only Jira Cloud is supported. The source editor also accepts an optional JQL filter.

Connecting Linear

Choose Linear and enter a personal API key with read access, a team UUID, and optionally a Linear project UUID. Open the team or project in Linear and use Cmd/Ctrl+K → Copy model UUID. Leave the key blank when editing to keep it; enter a new key to replace it.

Open includes triage, backlog, unstarted, and started issues; closed includes completed and canceled issues. All excludes archived issues. Search matches identifiers and titles among at most 1,000 recently updated issues in the configured scope. Issues show their status, assignee, and Linear link.

Custom HTTP sources

For anything else, the custom type points at any endpoint that returns a JSON array: configure the URL, request headers, and a field mapping that tells Shrimp Tide where an item's id, title, state and link live, as dot paths into the JSON.

Using sources while tracking time

In the task picker, each source shows up as its own group. Search matches titles and ticket ids, and you can filter by open or closed state and by assignee.

Permissions and security

• Configuring sources requires the permission to update the project — owners and managers.
• Browsing issues only requires access to the project's tasks.
• Tokens are stored server-side and always masked when read back; they never reach the browser.
• All provider requests are made by the server — your team's browsers never talk to Jira, GitHub or Linear directly.

Limitations

• Read-only: no write-back, no worklog sync, no status updates.
• Authentication is token-based; there is no OAuth flow.
• Provider errors and timeouts (after 10 seconds) appear inline in the task picker.