GitLab Issue Creation
Create GitLab issues from app feedback and operational events in the chosen project.
What it adds
A server-side path from an in-app report or system event to an issue in a configured GitLab project, on hosted or self-managed GitLab.
What your agent is told to do
5
What your agent is told to do
5-
1
Make the GitLab host part of the connection, not a constant. Self-managed installations live on customer-controlled domains, so validate the host, its reachability, and the credential's scope at setup, and store both server-side only.
-
2
Let an administrator map the destination project along with labels, milestone, assignee, and whether the issue is created confidential. Default to confidential when the report may contain customer data, and make that default explicit rather than implied.
-
3
Compose the issue on the server from the originating record, a deep link back into the app, and the acting user's identity, redacting credentials and personal data from any attached log before it leaves the app.
-
4
Key each submission by the originating record so a retry after a timeout finds the existing issue instead of creating a second one, and store the returned issue reference for later status reconciliation.
-
5
GitHub Issue Creation, Linear Issue Creation, Jira Ticket Creation, Trello Card Creation, and Asana Task Creation are siblings. The capture, redaction, deduplication, and reference storage belong to a single shared path; this brief owns only the GitLab host handling, field mapping, and error translation.
Edge cases it handles
8
Edge cases it handles
8- A self-managed host may sit behind a VPN, present an internal certificate, or run a version with different capabilities. Verify connectivity at setup, fail with a clear administrator-facing message, and do not assume the hosted service's behaviour applies.
- Labels, milestones, assignees, and the confidentiality flag must be selected from what the connected account can see in that project, and re-validated on use. A milestone closed or a label deleted since setup must not abort the submission; drop the unknown attribute, create the issue, and note what was dropped.
- Repeated submissions of the same report must converge on one issue. Fingerprint on the originating record or the error signature and comment on the existing issue rather than opening another.
- The project path, the host, and the token must never appear in client-side code, in an error message shown to a user, or in the issue body. A reporting user typically has no GitLab account at all and must not learn where private projects live.
- Projects get moved, renamed, transferred, or archived. Detect the redirect or the archived state, surface a configuration error to an administrator, and hold the queued reports rather than discarding them.
- Tokens expire and get revoked. Treat an authentication failure differently from a transient error: stop retrying, keep the submissions queued, and prompt an administrator to reconnect.
- Rate limits on hosted GitLab and on constrained self-managed instances require exponential backoff with jitter and a bounded retry count.
- When the GitLab instance is unreachable, accept the report in the app and create the issue later. The user sees their feedback recorded, not another organisation's outage.
Definition of done
9
Definition of done
9- The GitLab host is configurable and validated at setup, covering hosted and self-managed installations.
- An administrator chooses the project, labels, milestone, assignee, and confidentiality, selected from what the connection can see.
- Duplicate or retried submissions resolve to the existing issue rather than a new one.
- Host, project path, and token never appear in client-side code, user-facing errors, or issue content.
- A moved or archived project raises an administrator-facing configuration error while queued reports are preserved.
- Authentication failures stop retries and prompt a reconnection instead of dropping submissions.
- An unreachable instance still records the report, which is created once the host returns.
- The feature matches the existing design system.
- No existing functionality is broken.
Related features
AI Cost Budgets
AI Cost Budgets
Cap what AI features are allowed to spend before the bill arrives.
What it does
Monetary spending limits on AI work, scoped by workspace, feature, and time period, enforced before a run starts.
How it works
- 1 Find every place the app calls a model and route all of them through one accounting point that records estimated and actual spend against a scope. A budget that only covers the chat feature is not a budget.
- 2 Estimate the cost of a run from the size of its input before dispatching it, and refuse anything that would exceed the remaining budget on its own.
- 3 Reserve the estimate against the budget when the run starts, then reconcile to the real usage figures when it finishes, releasing whatever was over-reserved.
Copy the prompt
No account needed
Add this feature to my app:
https://addthisfeature.com/x/ai-cost-budgets
Prompt Versioning
Prompt Versioning
Tie every AI output to the exact prompt version that produced it.
What it does
Immutable, numbered versions of each prompt, with the run configuration recorded and every output stamped with the version used.
How it works
- 1 Make every publish create a new immutable version rather than overwriting the previous text. Editing history in place destroys the only record of what produced last month's outputs.
- 2 Capture the whole run configuration with each version, not just the wording: which model tier and parameters were used, which tools were available, and the expected output shape. A prompt that behaves differently under different settings is not one prompt.
- 3 Stamp every generated output with the version identifier that produced it, and keep that stamp with the record so an output found later can be traced back to its exact instructions.
Copy the prompt
No account needed
Add this feature to my app:
https://addthisfeature.com/x/prompt-versioning
Retrieval Debugger
Retrieval Debugger
Show exactly which sources, chunks, and scores produced a given AI answer.
What it does
A per-answer inspector showing the query as issued, the filters applied, the candidate chunks with their scores, and what reached the model.
How it works
- 1 Capture for each answer the query as it was issued, the filters applied, the candidates returned with their scores, and which of those actually made it into the request after the context ceiling was applied.
- 2 Show results after permission filtering, with a count of how many candidates were excluded and why. Displaying the pre-filter set turns the debugger into a way to read content the viewer cannot open.
- 3 Present each scoring stage separately — keyword, semantic, and any reranking — because a chunk that ends up first overall may have been rescued by one stage after being buried by another, and a single blended number hides that.
Copy the prompt
No account needed
Add this feature to my app:
https://addthisfeature.com/x/retrieval-debugger
How it works
-
1
Copy the link
Grab the Markdown instruction URL for this feature.
-
2
Give it to your AI
Paste it into Claude Code, Cursor, v0, Lovable — whatever you build with.
-
3
It inspects, then implements
Your agent reads your existing app first, then adds the feature to fit it.
Works with your stack
These instructions are written to adapt. They tell the agent to detect your framework, match your existing design system, and reuse what you already have — rather than assuming a particular stack.
Need it tighter than that? Customize the feature and tell it exactly what you're running.