Resources · Claude connector

Claude Qase Test Management connectorWhat Claude can do in your Qase workspace.

The Claude Qase Test Management connector exposes 13 tools: 5 read, 7 write, 1 undocumented. Published by Qase, this desktop extension lets Claude search your test repository, draft and update test cases, set up runs, record results and log defects.

Verified Trustpilot reviews · AI, automation & growth agency

Overview

What changes when Claude works in Qase

Qase holds your test cases, suites, runs and defects, and keeping them current takes a lot of clicking between screens. With this extension, you describe what you need and Claude does the work in the conversation: it reads the project, finds the right cases with Qase's own query language, then creates or updates what you asked for.

Turn a spec into test cases. qase_project_context shows the project's suites, milestones and environments, then qase_case_upsert creates each new case or updates an existing one.

Prepare a regression campaign in one go. qase_regression_run opens a run and adds every case from the chosen suites or plan, instead of picking them one by one.

Report a CI pipeline's results. From a finished pipeline's results, qase_ci_report creates the run, records every result and closes it; a failure worth tracking becomes a defect through qase_triage_defect.

What it will not do: nothing starts on its own when a build fails or a run ends; Claude acts only while it answers you, in the desktop app where the extension is installed. Defects cannot be linked to runs or results through Qase's API either. We have no n8n, Make or Zapier page for Qase; our Integrations hub lists the platforms we cover, and the Claude connectors page gathers the others.

Vocabulary

Five words before you install

The Claude vocabulary this page uses, in one minute.

Connector
The link you set up once between Claude and an account you already own, so Claude can work in it while it answers you.
Tool
One named thing a connector lets Claude do in that account. Claude chooses the tools it needs, and the directory sheet lists them by name.
Authorization
The access you give Claude through the service's own sign-in or permission screen, granted once per person and revocable the same way.
Approval
The confirmation Claude waits for before it goes ahead with something that changes the account, shown in the chat when it matters.
MCP
The common standard connectors are built on: it is what lets an assistant like Claude talk to an outside service such as Qase.
Connect

Install the Claude Qase Test Management connector in three steps

  1. 01

    Find Qase in the connector list

    In Claude Desktop, open the settings, go to Customize then Connectors, and search for Qase Test Management. On a Team or Enterprise workspace, an Owner or Primary Owner has to enable it for the organization before each member can install it.

  2. 02

    Start the install and sign in

    Launch it from its row with Connect and complete the steps Qase asks for to reach your workspace. If the link breaks later, Disconnect then install again: that is the documented way to repair it.

  3. 03

    Read what you are granting

    Before you confirm, read the access screen: it is Qase's screen, not Claude's, and it decides what the access covers. Claude then works within those rights and never beyond what your own Qase account can reach.

Tools

The 13 tools of the Qase Test Management connector

Qase Test Management gives Claude 13 tools: 5 that read your account, 7 that change something in it, and 1 no official source describes.

Five tools read your workspace, seven change it, and one is not classified. Names stay exactly as Claude displays them.

  • 5 read
  • 7 write
  • 1 not documented

What Claude reads (5)

5 tools

Five tools to explore a project, fetch any record and search across projects.

qase_discover_tools

Searches Qase's catalogue of additional tools, for needs such as deletions, test plans, milestone management or case reviews, and activates the ones it finds for the session. By default, only the core tools are visible.

When it helps
a request falls outside the thirteen tools on the directory sheet.
Watch out
Qase says a client that ignores its notice will not offer the new tool until the session is reconnected.

Sourcegithub.com · October 1, 2026 ↗

qase_get

Retrieves one Qase entity by its type and ID, such as a case, suite, run, result, plan, defect, milestone or environment, and can return only the fields requested. A project code is needed for entities that belong to a project.

When it helps
Claude has an ID and needs the full record behind it before answering.

Sourcegithub.com · October 1, 2026 ↗

qase_project_context

Returns a project's details, suite tree, milestones, environments, custom fields and users in one call. Qase recommends it as the first step when starting work on a project.

When it helps
Claude needs to know where suites live before filing new cases.
Watch out
each list stops at the first 100 items unless the full view is requested.

Sourcegithub.com · October 1, 2026 ↗

qql_help

Fetches one section of the Qase Query Language reference. A topic must always be named, because the whole reference is large, and Qase advises reading the entities section before querying an unfamiliar entity.

When it helps
a search targets an entity whose field names Claude has not checked yet.
Watch out
date fields differ between entities: a run has a start and an end, a result only an end.

Sourcegithub.com · October 1, 2026 ↗

What Claude changes (7)

7 tools

Seven tools that add or update test cases, runs, results and defects.

qase_case_upsert

Approval: see the rule

Creates a test case, or updates an existing one when its ID is given. Fields such as priority, severity or type accept either labels like high or blocker or numeric IDs, and the server normalises them.

What Claude asks for
no source documents a confirmation here; the general rule applies.
When it helps
a new test case must be written, or an existing one corrected.

Sourcegithub.com · October 1, 2026 ↗

qase_ci_report

Approval: see the rule

Records a CI pipeline's results in one call: it creates a run, records all results and can complete the run. Batches above 200 are split automatically, up to 2000 results per report.

What Claude asks for
nothing in the sources describes a confirmation for it; refer to the general rule.
When it helps
a pipeline has finished and its results belong in Qase.
Watch out
after a partial failure, repeating the call would create a second run; the remaining results go through qase_result_record.

Sourcegithub.com · October 1, 2026 ↗

qase_defect_upsert

Approval: see the rule

Creates a defect, or updates an existing one when its ID is given, status changes included. Setting the status to resolved is how an existing defect gets resolved.

What Claude asks for
no source describes an approval for this tool; the general rule covers it.
When it helps
a defect needs a new status, or a defect must be logged without starting from a failed test.

Sourcegithub.com · October 1, 2026 ↗

qase_regression_run

Approval: see the rule

Sets up a regression run in one call: cases are selected by suite IDs, explicit case IDs or a plan ID, then the run is created with every matching case, instead of the usual sequence of separate steps.

What Claude asks for
no source documents a confirmation for it; the general approval rule applies.
When it helps
a regression pass needs a run filled from whole suites or from a plan.

Sourcegithub.com · October 1, 2026 ↗

qase_result_record

Approval: see the rule

Records one or several test results into a run; each result needs a status, and linking the case is recommended. The bulk endpoint takes at most 200 results per request, and a longer list is refused before anything is written.

What Claude asks for
no source describes a confirmation for this tool; refer to the general rule.
When it helps
the results of a run have to be entered in Qase.
Watch out
results are append-only, so a retried batch would be recorded twice.

Sourcegithub.com · October 1, 2026 ↗

qase_run_upsert

Approval: see the rule

Opens a new test run, or edits an existing one when its ID is given. Qase's documentation of this tool says nothing more about the fields it accepts, so this page does not list any.

What Claude asks for
nothing in the sources describes an approval here; the general rule covers it.
When it helps
a run must be opened or adjusted outside the regression and CI shortcuts.

Sourcegithub.com · October 1, 2026 ↗

qase_triage_defect

Approval: see the rule

Creates a defect from a test failure. Title, actual result and severity are all required by Qase's API. Since that API cannot attach runs or results to a defect, the failing results are referenced in the actual result text instead.

What Claude asks for
no source documents a confirmation for it; see the general rule.
When it helps
a failed test needs a defect to track it.

Sourcegithub.com · October 1, 2026 ↗

Undocumented (1)

1 tool

One tool that Qase describes as raw API access, without saying whether it reads or writes.

qase_api

Qase describes it as a direct call to its REST API for endpoints the other tools do not cover: an HTTP method, a path and an optional body, in JSON only, so no file upload. No source says whether a given call reads or changes data, so this page leaves it unclassified.

Watch out
whatever a call does, the general approval rule below covers it.
Approvals

What Claude asks before it acts

By default, Claude stops and asks before each action it carries out in an account on someone's behalf. Seven of the thirteen Qase tools change your workspace, a thirteenth gives raw API access, and no source describes a confirmation specific to any of them, so that default is what to expect.

On Team and Enterprise workspaces, owners decide whether a member may let certain actions through without being asked each time. They can also restrict what a connector may do for the whole organization, for instance keep reading and close writing, and no member can override that from their own account. Claude acts with the rights of the person who connected it and nothing more: what that person cannot see or edit in Qase stays out of Claude's reach too.

Plans

Which Claude plans include it

Anthropic's general rule: remote connectors are open to all users on Claude, Cowork, Claude Desktop and mobile, while desktop extensions such as this one install in Claude Desktop.

On Team and Enterprise, an Owner or Primary Owner enables a connector for the organization before members can use it. For this extension's current status, check its sheet in the official directory, linked from the button at the top of this page. No official source publishes plan availability connector by connector; that sheet remains the only place to verify it.

Limits

What the Qase Test Management connector does not do

Only Qase documents this extension, on the public repository of its server: neither the Claude help center nor Anthropic's developer docs have a page about it.

A connector is not an automation: Claude calls its tools during a reply, and nothing fires when a pipeline finishes or a test fails. The directory's verification badge is not a security audit, and Anthropic states on every sheet that it does not control the tools a publisher exposes. The thirteen tools on the sheet are a floor, not a ceiling: Qase's discovery tool can switch on others during a session.

Need help

Need help connecting Qase Test Management to Claude?

A person reads every message.

FAQ

Questions about the Claude Qase Test Management connector

01What can Claude do with the Qase Test Management connector?
Claude can read a Qase project's structure, fetch any case, run, defect or other record, and search across projects with Qase Query Language. It can also create or update test cases and defects, open runs including full regression runs, record results, report a whole CI pipeline in one call and turn a failing test into a defect. Everything happens in the conversation in Claude Desktop, so you describe the job and Claude carries out the steps in your workspace.
02Can Claude write, send or delete anything through it?
Yes, seven of the thirteen tools change data in Qase: they create or update test cases, runs and defects, record results, report CI outcomes and set up regression runs. A thirteenth, qase_api, gives raw access to Qase's API, and no source says whether its calls read or write. Deletions are not among the thirteen listed tools, but Qase's discovery tool can activate extra ones, including deletions, during a session. So treat the list as a floor rather than a ceiling.
03Does Claude ask before changing my Qase workspace?
By default, Claude stops and asks before each action it carries out in an account on someone's behalf. No source describes a confirmation tied to any of the Qase tools that change data, nor to qase_api, so that general rule is what applies. On Team and Enterprise workspaces, owners decide whether members may let some actions through without a new prompt, and they can restrict what the connector is allowed to do for everyone in the organization.
04Which Claude plans is it available on?
No official source publishes, plan by plan, which connectors are included. Anthropic's general rule is that desktop extensions install in Claude Desktop for all users, and that on Team and Enterprise an Owner or Primary Owner enables a connector before members can use it. The connector's sheet in the official directory is the only place showing its current status, so check it from your own account. Being a desktop extension, it is used from the Claude Desktop app.
05Does Claude see my whole Qase account?
Claude works with the rights of the person who connected it and nothing more. Whatever that person cannot open or edit in Qase, Claude cannot reach either. Within those rights it can read projects, suites, cases, runs, results, defects, milestones and users, and search across projects. On Team and Enterprise workspaces, an owner can also narrow what the connector may do, for example reading only, for every member of the organization.
06Why does Claude refuse to record more than 200 results at once?
Qase's results endpoint accepts at most 200 results per request, and a longer list is refused before anything is written. It is not split automatically because results are append-only: a half-failed batch could not be undone, and retrying it would count the first part twice. For a finished pipeline, the CI report tool does the splitting itself, up to 2000 results in one report, and beyond that the suite is reported as several runs.
07Claude or an automation tool for Qase?
They do different jobs. This connector serves a conversation: you ask Claude to draft cases, prepare a run or log a defect, and it does so while you talk. An automation tool runs a fixed scenario by itself when an event happens. Nothing in this extension starts on its own, so pushing results automatically with nobody asking is a job for an automation tool, while Claude works when you ask it to.