mirror of https://github.com/microsoft/PowerToys.git synced 2026-09-01 19:51:34 +02:00

Files

Niels Laute ddc536c696 fix(ci): correct issue triage labels and repro detection (#49999 )

## Summary of the Pull Request

Fixes two automated issue-triage regressions exposed by #49989:

- Stops `.github/workflows/issue-triage.md` from adding, removing, or
replacing version labels. Reported versions remain available for the
triage summary and update guidance only.
- Expands `.github/scripts/issue-triage/issue-context.py` action
detection so concise natural-language reproduction steps using verbs
such as `make`, `hold`, and `use` are treated as actionable.
- Adds regression coverage for #49989 and a workflow contract test that
prohibits version-label management.

## PR Checklist

- [x] **Communication:** Discussed with a core contributor based on the
automation behavior observed in #49989
- [x] **Tests:** Added/updated and all pass
- [x] **Dev docs:** Updated `.github/scripts/issue-triage/README.md`

## Detailed Description of the Pull Request / Additional comments

The issue body in #49989 was not edited after creation. The workflow
received the original reproduction steps, but deterministic
preprocessing recognized only one action verb (`press`) and classified
the steps as insufficient because two action markers were required. The
publisher then added `Needs-Author-Feedback` from that result.

Independently, the publisher explicitly matched the reported PowerToys
version against repository labels and added `0.100.2`. That
version-label mutation has been removed from both the source workflow
and generated lockfile, with a static contract test to prevent it from
returning.

This PR does not close #49989 because that issue tracks the underlying
Keyboard Manager behavior, not the triage automation regression.

## Validation Steps Performed

- `python -m unittest discover .github\scripts\issue-triage\tests` — all
50 tests passed.
- Evaluated the exact #49989 issue body through `reproduction_quality`;
it now returns `SUFFICIENT`.
- Regenerated `.github/workflows/issue-triage.lock.yml` with `gh aw
compile issue-triage --no-check-update` using gh-aw v0.86.2.

Copilot-Session: 498721d4-1098-4298-ac8a-147066fbfea3

2026-08-19 16:44:21 +02:00

tests

fix(ci): correct issue triage labels and repro detection (#49999 )

2026-08-19 16:44:21 +02:00

bug-report-analyzer.py

…

issue-context.py

fix(ci): correct issue triage labels and repro detection (#49999 )

2026-08-19 16:44:21 +02:00

README.md

fix(ci): correct issue triage labels and repro detection (#49999 )

2026-08-19 16:44:21 +02:00

verify-agent-output.py

Improve issue triage product-label detection (#49905 )

2026-08-14 21:14:55 +00:00

README.md

AI-assisted issue triage

The workflow in .github/workflows/issue-triage.md maintains one canonical triage comment when an issue is opened or its original title/body is edited. It combines deterministic preprocessing with one bounded GitHub Copilot pass.

Rules

Parse the issue template, PowerToys version, product area, reproduction quality, language, and diagnostic-report requirement before AI runs.
Compare the reported version with the latest stable PowerToys release. Older versions receive a non-blocking update-and-retest recommendation.
Retrieve a bounded set of older candidate issues using product labels, technical identifiers, and focused title/body searches.
Ask Copilot only to summarize the issue, judge supplied duplicate candidates, interpret sanitized diagnostics, and classify the author-written language.
Maintain one marked comment with separate sections for the issue author and the PowerToys team.
Mention the author once and list only needed or recommended actions.
Apply Needs-Author-Feedback when blocking information or an English translation is required. Removing the label disables scheduled closure.
Add a matching primary Product-* label without removing existing product or maintainer labels.
Never add, remove, or otherwise manage version labels.
Submit duplicate closure as a native GitHub suggestion. A maintainer must accept or decline it; acceptance closes the issue as a duplicate and links it to the selected canonical issue.
Never close an issue directly from the model output.

Reproduction and diagnostics

Concrete actions plus an observed result are sufficient.
Passive or intermittent failures can be sufficient when the timing/trigger and observed failure are clear.
Non-English reproduction steps are reassessed after translation rather than treated as missing.
Diagnostic reports are required for crashes, hangs, startup/load failures, installation/update failures, performance failures, and system-integration failures.
Reports are optional for clear UI/visual defects and recommended for other actionable bugs.
Report ZIP files are validated for path traversal, encryption, archive size, file count, and decompressed size. Only bounded redacted evidence reaches Copilot; raw archives are deleted and never uploaded as artifacts.

Author-feedback lifecycle

.github/policies/resourceManagement.yml closes open issues and pull requests that retain Needs-Author-Feedback for seven days without activity.

An author comment removes Needs-Author-Feedback and returns the item to team triage.
A push to a non-draft PR reruns PR intake, which recalculates Needs-Author-Feedback.
Manual label removal immediately makes the item ineligible for scheduled closure.

Cost and safety controls

The small model alias is limited to five turns and 10 AI credits per run.
The workflow subscribes only to issue creation and edits to the original issue; comments and reopen events do not trigger it.
Per-user rate limits, daily AI-credit limits, and per-issue concurrency bound repeated issue creation or edits.
The agent has no general shell or GitHub API tools. Its only shell command is the structured safe-output CLI proxy, and Copilot's file-write tool is explicitly denied to work around gh-aw v0.86.2 treating edit: false as writable.
Threat detection fails closed; publication requires an explicit successful detection result.
The publishing job rebuilds evidence from the current issue and accepts only deterministic product-label candidates, duplicate candidates, hashes, and classifications. Stale or manipulated model output fails before any write.
A separate validated safe-output job owns comment, label, and duplicate-suggestion writes.

Retired automation

The GitHub Models-based automatic issue deduplicator is removed.
The GitHub Models-based issue/PR area labeler is removed. This workflow replaces issue labeling only; automatic PR product labeling is intentionally not replaced here.
The Azure Pipelines XAML Styler verification step is removed. The local .pipelines/applyXamlStyling.ps1 developer tool remains available.

Run the focused tests with:

python -m unittest discover .github\scripts\issue-triage\tests -v