Repository navigation
Description
Motivation & Background
While performing backlog triage against main git history and merged pull requests, I identified several open issues whose requested implementations and tests have already landed on main. These remained open because the corresponding PR descriptions cited Refs #XXXX rather than GitHub's closing keywords (Fixes #XXXX / Closes #XXXX).
To ensure our issue tracker accurately reflects the current state of the codebase, I conducted an evidence-backed audit across open issues. Below is a breakdown of 10 fully resolved issues pending close-out, along with 2 stale proposals for maintainer direction.
Before proceeding with individual issue closures, I wanted to share this overview with the team (@potiuk, @justinmclean) to confirm alignment.
1. Resolved Issues Pending Close-out
The following 10 issues have their complete requested implementations and test suites merged into main:
| Issue | Title | Merged PR(s) & Commit(s) | Verification / What Landed on main |
|---|---|---|---|
| #1655 | Welcome guide: pr-management family | PR #1701 (76de43ab) |
Added plugins/magpie-setup/skills/setup/setup_preflight/guides/pr-management.md and test coverage in tools/setup-preflight/tests/test_welcome.py. |
| #1654 | Welcome guides for every skill family | PRs #1653, #1669, #1701, #1702, #1706, #1708, #1709, #1724, #1725, #1726 | All 10 skill family welcome guides are merged in plugins/magpie-setup/skills/setup/setup_preflight/guides/ and verified by test suite. |
| #1069 | Add a proper Mailing List page | PR #1629 (b298450a) |
Added docs/mailing-lists.md with subscription and archive details, linked from docs/index.md. |
| #303 | feat(tools/mail-source/imap): land concrete CLI wiring for the IMAP stub | PR #1466 (3c86798e) |
Landed concrete CLI under tools/mail-source/imap/ covering contract operations (read_thread, list_recent_threads, create_draft, etc.) with test suite. |
| #1924 | imap: read_thread rejects a bare Message-ID the contract says it accepts | PR #1962 (d5dd37f0) |
Added canonicalisation of bare Message-IDs across all thread operations with regression tests. |
| #1824 | hook-manager: add a CaptainHook adapter | PR #1899 (7ca3873a) |
Added CaptainHook adapter to tools/hook-manager/registry.json, updated shim, documentation, and tests. |
| #1416 | feat(adapter/grok): xAI Grok agent harness and adversarial-review backend | PRs #1446 (3e116089), #1488, #1516 (faa87668), #1593, #1598 |
All planned Grok integration points (reviewer backend, setup recognition, PreToolUse guard, spec-loop profile, sandbox-lint declarations) merged. |
| #321 | feat(adapter/junie): JetBrains Junie skill runtime | PR #1666 (6c24f30c) |
Added docs/adapters/junie.md, --junie adapter for agent-guard, registry entries, and test coverage. |
| #76 | Auto-skip when a triager has already left a similar comment | PR #1420 (7eb89add) |
Multi-triager marker recognition (OWNER/MEMBER/COLLABORATOR), by= field in pr-triage-fold, and attribution in triage reason templates. |
| #1173 | release-verify-rc: validate rc jars | PRs #1415 (9c075618), #1505 (78032a90), #1506 (3d4054b6), #1619 (abd59697) |
Checks 1–7 (POM ALv2, incubator disclaimer, sources/javadoc presence, Nexus staging, reproducibility/namespace) merged and classified as code. |
2. Stale / Superseded Proposals for Review
- feat(tools/pagure): tracker bridge for Fedora and pagure.io-hosted projects #312 (
feat(tools/pagure): tracker bridge for Fedora and pagure.io-hosted projects): The upstream public instancepagure.iohas sunsetted and returns 404. Should we close this asnot planned, or is support for self-hosted Pagure instances still desired? - Consider delegating the packaging layer to microsoft/apm #518 (
Consider delegating the packaging layer to microsoft/apm): Discussion concluded without an active POC, and the author confirmed they will not be pursuing this. Magpie's distribution architecture (in-tree plugins and marketplace) has established its own path. Can this be closed as superseded?
Proposed Next Steps
If maintainers are in agreement:
- I will close the 10 resolved issues in Section 1 with individual closing comments citing the respective PR and commit SHAs.
- We can resolve or close the two stale proposals based on community feedback.
- Once completed, I will close out this tracking issue.
Feedback welcome!
Activity
Kaap10 commentedon Nov 14, 2026
@shahar1 can you please have a look on this, so that we can close these stuffs right away?
choo121600 commentedon Nov 15, 2026
I've reviewed the issues, and Section 1 looks resolved to me. I think the remaining issues could potentially be closed as well, however, I'm not very familiar with APM, so I'd like to hear what others think before we make a final decision.
Kaap10 commentedon Nov 15, 2026
Thank you @choo121600. All 10 resolved issues in Section 1 have now been closed with references to their respective merged PRs and commits.
Keeping this tracking issue open while awaiting feedback from other regarding the two proposals in Section 2.
feat(utilities): add pre-release-sweep, the consistency sweep before …
feat(utilities): add pre-release-sweep, the consistency sweep before …
potiuk commentedon Nov 15, 2026
Thanks @Kaap10 — this is a useful audit, and checking
mainrather than the issue labels is the right way to do it.I re-verified the ten closures item by item against
main: none was wrong.Eight are fully resolved (#1655, #1654, #1069, #303, #1924, #1824, #321, #76).
Two closed with small leftovers, which I'll open as follow-up issues rather than reopen:
tools/privacy-llm/models.mdor a recipe indocs/setup/privacy-llm.md(the existing mention covers Grok only as an adversarial-review CLI, which is ungated by design).META-INF/LICENSE/NOTICEinside the jars.On Section 2:
tools/forgejo/already covers Forgejo / Gitea. A self-hosted Pagure bridge can come back if an adopter needs one.Most of these stayed open because the PRs said
Refs #Ninstead ofFixes #N/Closes #N, so please use the closing keywords when a PR finishes an issue. #2110 adds apre-release-sweepskill that runs this audit before each release; its eval cases are taken from this issue.With that, this issue can be closed.