Of the 623 acted-on closes, 71 were likely fixed
instead — closed as completed without the sweeper's comment on a proposal other
than implemented-on-main. Most plausibly the team fixed the issue in the meantime; these are
not claimed as adoption (blue in the acceptance chart below).
2nd-reviewed before the close: 100 confirmed · 33 disputed · 4 unverifiable · 486 never 2nd-reviewed — the verify lanes are newer than most of these closes; the split fills in from here. Details in Independent verification.
Only closures of issues the sweeper proposed to close. Weeks start Monday.
Each bar is one close reason — the same merged rows as the verdicts page's verdict distribution (structured reason first, legacy close labels folded in; the label-free close is implemented on main). Right label: acted-on / proposals. "Closed otherwise" counts closures without the sweeper's comment — correlation, not proven adoption; blue splits out the likely-fixed-instead closes, which the sweeper doesn't claim at all. Newer proposals have had less time to be acted on, so compare reasons with similar proposal ages cautiously.
70 of the 303 own-wording closes were likely fixed instead (closed as completed) — those proposed comments weren't passed over for a better one; the close had a different cause entirely.
Draft PRs opened with the sweeper-fix skill,
discovered by reconcile's seeded-by-marker search — per user, against their
agent-fixable pool (the track). Corpus-wide pool: 1,771 reviewed
fix specs. As current as the last reconcile run.
A blinded second review re-judges duplicate and implemented-on-main close proposals before anyone acts (the dashboard's "2nd review" badge). "Disputed" is a dispute to read, not a proven error — the blinded judge sees less context than the original review. Coverage counts open proposals (the verify sweep's selection universe); stale stamps (the record was re-reviewed since verification) are excluded from the mix, and live issue drift is re-checked by the nightly verify sweep, not this page. Scope is these two close reasons only.
duplicate: 620 of 620 open proposals verified · 107 closed unverified — implemented-on-main: 526 of 526 open proposals verified · 97 closed unverified
Duplicates by the canonical's kind, implemented-on-main by claim kind (hover a row label for the lane). Right label: disputed / verified — a kind disputing well above its peers is a calibration signal for the review prompt.
The 389 disputed proposals are listed on the dashboard (the link lands pre-filtered); each ✗ badge carries the verifier's evidence. A dispute is worth a manual look before acting: either the review or the verifier is wrong, and both readings are on record.
Counts, not rates — closed-after-verification volume is still small, so percentages would mislead. A disputed proposal closed anyway and later reopened is a false positive the verify lane flagged in advance.
| 2nd-review verdict | Still open | Closed since | Reopened after a close |
|---|---|---|---|
| confirmed | 760 | 99 | 1 |
| disputed | 356 | 33 | 0 |
| unverifiable | 30 | 4 | 0 |
Closes that went ahead after the 2nd review refuted the proposal — either the human knew better (the dispute is advisory) or a false positive slipped past the warning; both readings are on record. The before-the-fact twin of the reversals queue below.
| Issue | Closed | Who acted | 2nd-review evidence |
|---|---|---|---|
| #277300 Google Authentication stopped working in 1.106 | 2026-08-31 | chrmarti | A is a concrete 1.106 regression breaking Google OAuth token refresh with 'unable to get issuer certificate'; B only tracks rollout/telemetry of the experimental http.systemCertificatesNode setting, and completing it needn't fix A. |
| #130172 Proxy Credentials Popup not showing | 2026-08-31 | chrmarti | Commit adds proxy basic-auth lookup for extension-host requests via proxy-agent, but node and browser RequestService — the path settings sync uses — explicitly return undefined, so the reported missing sync credentials prompt persists. |
| #292795 "Invalid String Length" in Copilot Chat | 2026-08-31 | chrmarti | A reports 'Invalid String Length' with no images, attachments, or tool output involved and says model/mode is irrelevant; B's root cause is base64 image/binary payloads inlined into session JSONL — shared symptom, unestablished shared trigger. |
| #299324 Sorry, there was a network error. Please try again later. Request id: 836774d3-2064-4706-8d01-947d2a4d2cd0 Reason: Please check your firewall rules and network connection then try again. Error Code: net::ERR_HTTP2_PROTOCOL_ERROR: [object Object]. | 2026-08-31 | chrmarti | Both are symptom-only auto-reports of a generic ERR_HTTP2_PROTOCOL_ERROR on different OSes (Linux vs Windows) and extension versions; an environment-dependent network/proxy failure string alone establishes no shared root cause. |
| #300513 CoPilot Chat 0.39.0 is incompatible with latest VSCode 1.111.0 | 2026-08-31 | chrmarti | Different release pairings a cycle apart — B is Copilot Chat 0.38.0 against VS Code 1.110, A is 0.39.0 against 1.111 — so resolving B's version-gate mismatch would not cover A's later recurrence. |
| #301098 When reopening VS Code, recent chat sessions can’t be opened. | 2026-08-31 | chrmarti | Both share the symptom of a chat session not opening, but A's is transient after restarting a remote SSH window and self-heals in days, while B's persists for one specific local session across reinstall/re-auth. |
| #303402 Remote SSH workspace paths are handled locally in Copilot Chat UI mode, causing repeated tool failures and renderer overload | 2026-08-31 | chrmarti | B is Windows-host backslash mangling of the workspace path causing 'outside of workspace' errors and false-success writes; A is macOS-host ENOENT on correct POSIX paths plus a retry loop overloading the renderer. |
| #328438 Generate Commit Message not working | 2026-08-29 | dmitrivMS | Both are bare 'commit message generation produces nothing' reports on different builds (1.131.0 vs 1.130.0) with no logs, errors, or Copilot state, so no shared root cause is established beyond an identical symptom. |
| #329339 Commit generation personalized instruction doesn't work | 2026-08-29 | dmitrivMS | B fails with promptFiltered specifically on Japanese-language instruction text; A fails with an unprocessable-entity error on a plain ASCII English instruction — different failure signatures and triggers, not one shared root cause. |
| #330831 Modern UI: Some panel parts use Sidebar colors instead of Panel colors | 2026-08-28 | na2co3-ftw | The commit adds a chat customization migration hint (settings, command, storage key); it touches no paneHeader.css or modern-UI theming code, so it cannot address panel parts using sideBar color tokens. |
| #327578 Chats system model: New chat button position changes, resulting into annoying behavior | 2026-08-27 | romalpani | gate: the cited canonical #329321 is closed — a duplicate close would strand the report (re-review, not close) |
| #312110 PendingMigrationError: navigator is now a global in nodejs thrown during module load on remote VS Code server, causing repeated renderer crashes and Remote SSH disconnections | 2026-08-27 | chrmarti | Same PendingMigrationError signature, but different offending code: A is the bundled Copilot extension touching navigator at top-level module load on the remote server; B is the Remote-SSH extension's own bundle locally — separate fixes. |
| #331641 J0t.clearMarks is not a function | 2026-08-27 | vs-code-engineering | Only a shared minified symptom string; neither report gives a trigger or stack, and they run different extension/VS Code builds (0.35.3/1.107.1 vs 0.48.1/1.111.0), so identical minified 'J0t' does not establish one root cause. |
| #326199 VS Code does not connect to tunnel after updating | 2026-08-27 | alexdima | A fails with an explicit 'Client refused: version mismatch' handshake rejection after a client-only update, whereas B updated the tunnel service first and fails with an unexplained WebSocket 1006 close — different failure mechanisms. |
| #231128 font ligatures that span several textmate scopes assume the font settings of last character (last component of the composite symbol) | 2026-08-27 | alexdima | B asks for a setting to suppress ligatures inside constructs like strings for readability; A reports miscolored glyphs when a ligature straddles two scopes — implementing B's opt-out wouldn't fix A's tokenization/coloring defect. |
| #329741 Black Screen after 10 minutes | 2026-08-27 | deepak1556 | gate: the cited canonical #261660 is closed — a duplicate close would strand the report (re-review, not close) |
| #323592 Agents window: "New Chat" button is a no-op from within a current session | 2026-08-26 | justschen | B documents debounce/queueing flakiness from rapid clicks plus a no-op when already on an empty New Chat; A is a no-op while inside an existing active session — different trigger state, not a shown shared cause. |
| #327257 Copilot Chat: file links render as broken vscode-file:// URIs instead of workspace paths | 2026-08-26 | justschen | A reports backtick content resolving to a malformed vscode-file://…/workbench.html app URI that navigates nowhere; B requests replacing optimistic symbol linkification with explicit path anchors — a design change, not that URI-resolution defect. |
| #295956 Right-to-Left Text Direction for some languages | 2026-08-26 | justschen | gate: open PR #314365 references the issue — someone may be fixing it; closing as duplicate is not safe |
| #295731 Preserve pasted screenshots in Copilot chat across IDE restarts / session reloads | 2026-08-26 | justschen | Commit only persists attachments of the empty/untitled draft input (chatSessionIsEmpty) via a separate storage memento; it does not restore images in an existing chat session's transcript, nor add the requested per-chat clear/disable persistence settings. |
| #288926 "Retry" removes attached context | 2026-08-26 | justschen | The diff only hoists accessor.get calls before awaits to avoid a disposed-accessor error in the retry action; it never touches request variables/attachments, so the reported loss of pinned context on retry is unaddressed. |
| #288574 Expanding thinking block scrolls down a bunch, scroll should be stable | 2026-08-26 | justschen | The diff only retimes auto-scroll-to-bottom inside the thinking part's own scrollable during streaming; it never anchors the outer chat list's scroll position when the user toggles expansion, the reported trigger. |
| #286231 Terminal inline chat has `Unknown variable type` reference | 2026-08-26 | justschen | The diff only appends the variable name to the fallback label for diagnostics; the reference still falls into the 'Unknown variable type' branch, so the reported inline-chat rendering defect remains. |
| #284142 Chat View Lacks Smooth Scrolling for Long Responses | 2026-08-26 | justschen | A asks for manual navigation fixes (tiny scrollbar, arrow keys jumping between prompt turns instead of scrolling); B reports auto-scroll during streaming mispositioning and stalling — different mechanisms, fixing B leaves A's keyboard/scrollbar problems. |
| #282664 "Start Impl" btn seems home grown | 2026-08-26 | justschen | Commit only re-plumbs plan approval to a vscode_reviewPlan tool and renames action labels; no change to the split-button's rendering, alignment, or hover-target width the report describes. |
| #280525 Chat response thinking is not aligning vertically in the center | 2026-08-26 | justschen | Diff is a broad thinking-UI redesign (fade masks, connector, left padding, scroll caching); nothing addresses the reported asymmetric top/bottom spacing from markdown paragraph bottom margins in the collapsed preview line. |
| #265795 Please Fix Shared Input and History Across Multiple Copilot Chat Sessions | 2026-08-26 | justschen | Commit only scopes up/down input history per session id; it does nothing about the report's primary defect that typed input appears simultaneously in all chat sessions. |
| #253296 Slight vertical jump when editing chat message | 2026-08-26 | justschen | Commit fixes an input-disposal bug and removes a hover-offset rule; neither addresses the ~5px vertical shift of a chat message when entering edit mode as reported. |
| #252501 Clicking to edit a copilot message flickers for ~1 frame | 2026-08-26 | justschen | Diff guards focus-outside firing while another request is already being edited and fixes inputContainer teardown on finishing edit; neither covers the reported one-frame flicker when first clicking a message to edit. |
| #306236 Incorrect Minimap Vertical Size | 2026-08-25 | alexdima | A reports minimap 'fit' mis-sizing constantly on any tab switch with no repro; B requires a specific trigger chain — opening then closing the terminal panel from a larger file — so the defects aren't clearly the same. |
| #236365 "Undo" history lost after "Save as". | 2026-08-25 | alexdima | A's body describes a plain Ctrl+S on a 180MB CSV leaving a degenerate undo stack (blank file ↔ final state), not B's Save As discarding an otherwise intact undo/redo buffer. |
| #318485 Replying to a past conversation results in "Input item ID does not belong to this connection: | 2026-08-19 | vs-code-engineering | A is a deterministic conversation-state bug ('Input item ID does not belong to this connection' when replying to a past chat), reproducible across retries and models — outside B's stated scope of transient/service-side request failures. |
| #330728 fix: filter discovered pre-existing session in session config integration tests (build fix for vscode-engineering#3607) | 2026-08-16 | sandy081 | Commit serializes AgentPluginManager cache loading; it never touches sessionConfig.integrationTest.ts's root/sessionAdded predicate, so the discovered PRE_EXISTING_SESSION_URI notification race the issue reports remains unfiltered. |
Close proposals whose closure was later reversed (the issue reopened) — 1 so far, 0 of them after a close that used the sweeper's comment. A reversal of a sweeper-comment close is the strongest false-positive signal the pipeline produces — worth a manual look (and a feedback issue) each. A reopen isn't automatically a wrong verdict: issues also reopen to track follow-up work or get re-closed differently.
| Issue | Reopened | Reversed close used | Who had acted |
|---|---|---|---|
| #255890 When you activate a source code chunk it scrolls to the top. | 2026-07-28 · closed again since | own wording | justschen |
Attribution: poster of the matched comment, else who closed the issue.
| User | Closures | With sweeper comment |
|---|---|---|
| alexdima | 114 | 107 (94%) |
| justschen | 107 | 54 (50%) |
| vs-code-engineering | 50 | 0 (0%) |
| roblourens | 40 | 17 (43%) |
| hediet | 35 | 35 (100%) |
| lramos15 | 33 | 0 (0%) |
| kycutler | 30 | 14 (47%) |
| dmitrivMS | 26 | 0 (0%) |
| chrmarti | 18 | 17 (94%) |
| sbatten | 16 | 16 (100%) |
| benibenj | 15 | 10 (67%) |
| deepak1556 | 13 | 0 (0%) |
| aeschli | 12 | 11 (92%) |
| sandy081 | 12 | 0 (0%) |
| lszomoru | 9 | 8 (89%) |
| (unattributed) | 8 | 0 (0%) |
| mrleemurray | 7 | 5 (71%) |
| vritant24 | 7 | 3 (43%) |
| egamma | 6 | 6 (100%) |
| TylerLeonhardt | 6 | 6 (100%) |
| joshspicer | 5 | 0 (0%) |
| karthiknadig | 5 | 2 (40%) |
| pwang347 | 5 | 3 (60%) |
| bhavyaus | 4 | 4 (100%) |
| meganrogge | 4 | 0 (0%) |
| alexr00 | 3 | 0 (0%) |
| dbaeumer | 3 | 1 (33%) |
| benvillalobos | 2 | 0 (0%) |
| AbhaySanthani-tekframeworks | 1 | 0 (0%) |
| aiday-mar | 1 | 0 (0%) |
| AndrewStopchenko-SO | 1 | 0 (0%) |
| babakzarrinbal | 1 | 0 (0%) |
| BladeJoe | 1 | 0 (0%) |
| BobVul | 1 | 0 (0%) |
| bryanchen-d | 1 | 0 (0%) |
| coder-free | 1 | 0 (0%) |
| connor4312 | 1 | 0 (0%) |
| danwilhelm | 1 | 0 (0%) |
| DavidLangworthy | 1 | 0 (0%) |
| dileepyavan | 1 | 0 (0%) |
| DonJayamanne | 1 | 0 (0%) |
| dustintran333 | 1 | 0 (0%) |
| JMS-1 | 1 | 0 (0%) |
| jruales | 1 | 0 (0%) |
| Mapalmeira | 1 | 0 (0%) |
| monolithed | 1 | 0 (0%) |
| na2co3-ftw | 1 | 0 (0%) |
| Prasanna-2005 | 1 | 0 (0%) |
| romalpani | 1 | 0 (0%) |
| sanket-bhalerao | 1 | 0 (0%) |
| SimonSiefke | 1 | 0 (0%) |
| steven8274 | 1 | 0 (0%) |
| Tanishq-JM | 1 | 0 (0%) |
| trent-abc | 1 | 0 (0%) |
| ulugbekna | 1 | 0 (0%) |
| xgtxdzh | 1 | 0 (0%) |
Every reviewed issue that closed, by what the review had concluded. Only the propose-close row is claimed by the funnel above; needs-info issues auto-close on their own cadence, and keep-open closures happened for unrelated reasons (worth eyeballing as a precision check).
| Review verdict | Closed |
|---|---|
| keep-open | 697 |
| propose-close | 623 |
| needs-info | 422 |
| route-to-area | 117 |
| Close reason | Closed |
|---|---|
| completed | 942 |
| not_planned | 659 |
| duplicate | 247 |
| not_found | 11 |