/
/
Verify a GitHub issue fix with a fresh Ghost Inspector run
For developers who want issue closures backed by a new test run, recorded on the issue, by asking their AI assistant.
On this page
Connector
Connect GitHub, pick your AI tool and paste the prompt.
The problem
Issues closed on a hunch
Closing a GitHub issue about a test failure should rest on a fresh Ghost Inspector run, but finding the test, running it and recording the result takes several steps. Issues get closed on a hunch or sit open after the fix.
Without PopdockAI
- Read the issue to find which test it means
- Run the test in Ghost Inspector
- Copy the result back to GitHub
- Close the issue on a hunch
With PopdockAI
- The issue and comments are read to find one test
- The test runs once after you confirm
- The result and any failing step are posted as one comment
- A pass suggests closing, but the issue is never closed for you
The prompt
Use this prompt
Copy it into any AI tool connected to PopdockAI. Swap the values in braces for your own.
**Objective:** Check whether the failure described in GitHub issue `{issue-number}` is fixed by re-running its Ghost Inspector test (`{test-id}` if given, otherwise the test named in the issue), and record the result on the issue.
**Step 1: Read the issue (GitHub)**
Call `github-get-issue` with `issue-number` set to `{issue-number}`, and `github-get-issue-comment` with the same `issue-number`. If the issue is not found, stop.
**Step 2: Identify the test (Ghost Inspector)**
* If `{test-id}` is given, call `ghost-inspector-get-test` with `id` set to it.
* Otherwise look in the issue title, body and comments for a Ghost Inspector test id or test name (issues filed from Ghost Inspector list each failing test's name and test id). Call `ghost-inspector-get-test` with `id`, or with `name.contains` set to the name.
If no test can be identified, or more than one matches, list what you found and stop so the user can supply `{test-id}`. If the issue names several tests, list them and ask whether to run all of them or one.
**Step 3: Confirm and run (Ghost Inspector)**
Tell the user which test you will run (name, suite, start url) and wait for confirmation. Call `ghost-inspector-execute-test` with `test-id` set to the test's id. Do not set `immediate`.
If it fails, call `ghost-inspector-get-test-step` with `test-result-id` set to the new result id and `passing` false to get the failing step.
**Step 4: Record it (GitHub)**
Call `github-add-issue-comment` with `issue-number` `{issue-number}` and a `body` stating the test name, the new result id, pass or fail, execution time, and the failing step's command and target if it failed.
**Step 5: Report**
Pass or fail, and if it passed, suggest the issue can be closed. Do not close it yourself.
**Rules & Guidelines:**
* The github tools are on the github server; the ghost-inspector tools are on the ghost-inspector server.
* Writes are exactly one `ghost-inspector-execute-test` per confirmed test and one `github-add-issue-comment`.
* Do NOT call `github-update-issue` (that would close or edit the issue), `ghost-inspector-accept-screenshot` or `ghost-inspector-update-test`.
**Validation:**
* If the issue is already closed, say so and ask whether to run the test anyway.
* A single pass after a history of intermittent failures is not proof of a fix. If the issue or its comments call the test flaky, say that one pass is weak evidence.Who it's for
One answer, three teams that use it
Developers
Prove your fix before closing
A fresh run and its result are posted on the issue, so reviewers see the evidence.
Verify the fix for GitHub issue 388 by running its Ghost Inspector test.
QA engineers
Record test results on the issue
The result, with the failing step if it failed, is posted as one comment.
Run Ghost Inspector test 5f2a91 for GitHub issue 140 and post the result.
Maintainers
Close issues with confidence
A pass suggests the issue can close, with a warning when the test is known to be flaky.
Is GitHub issue 75 in northwind/web really fixed? Rerun the related Ghost Inspector test.
How it works
What PopdockAI does when you ask
You give the issue number and optionally a test id, and your AI assistant verifies the fix with one fresh run. It warns that one pass is weak evidence for a flaky test.
01
Ask
You give the GitHub issue number and optionally a test ID.
02
Find
It reads the issue and comments to identify one Ghost Inspector test.
03
Confirm
You confirm the test and it runs once.
04
Record
It posts the result as one comment and suggests closing on a pass.
What PopdockAI provides
The tools behind the answer
Ready made tools for GitHub and one secure connection to your AI tool. No code, no exports.
github-get-issue
github-get-issue-comment
ghost-inspector-get-test
ghost-inspector-execute-test
ghost-inspector-get-test-step
github-add-issue-comment
This report is built from Get data tools only. PopdockAI applies the row, field and parameter rules your admin sets, and every query is written to the audit trail.
The outcome
Fixes verified, with the evidence on the issue
Related use cases
More ways to put this data to work
One prompt template with infinite possibilities.
Your AI assistant adapts its approach, tone, and capabilities to match each unique context while you maintain a single, manageable workflow.
Join forward-thinking companies that are transforming their operations with intelligent, secure AI automation.

.webp)
