[pull] main from microsoft:main - #1611
Merged
Merged
Conversation
…rts (#331709) * agentHost: avoid collecting Copilot process logs Disable SDK process-log collection during debug exports while preserving session event and shell logs. Add focused coverage for both session-scoped and host-wide requests.\n\n(Written by Copilot)\n\nCo-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * agentHost: address debug log review feedback Document why SDK process logs are excluded and make the debug-log mock return destination-specific paths. (Written by Copilot) Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * agentHost: remove redundant process log comment (Written by Copilot) Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> --------- Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
* base: share WebSocket framing across tunnel transports Extracts the RFC 6455 frame parser and encoder into the shared IPC layer so node IPC sockets and browser tunnel connections use the same implementation. - Supports masked client frames, fragmented messages, extended payload lengths, control frames, close handshakes, and protocol validation. - Removes the websocket framing package and its browser compatibility shims from the Dev Tunnels web bundle. - Adds focused coverage for the shared codec and browser tunnel transport. (Commit message generated by Copilot) Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * base: harden shared WebSocket framing Addresses review feedback for lifecycle, resource limits, and high-throughput IPC performance in the shared RFC 6455 implementation. - Restores browser client frame, message, and close-handshake limits. - Adds zero-copy in-place unmasking for the Node IPC path. - Handles Pong frames explicitly and adds focused regression coverage. (Commit message generated by Copilot) Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> --------- Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
* agentHost: improve debug log export Show collection progress, include rotated VS Code logs, and use one 256 MiB artifact limit.\n\n(Written by Copilot)\n\nCo-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * agentHost: harden debug log export Snapshot rotated logs, use stable paths, reject symlinks, and enforce limits across the complete ZIP.\n\n(Written by Copilot)\n\nCo-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
* Tweak builtin pr skills to not ban gh cli If the gh CLI is available and authed, some tools are not enabled in the gh mcp * Potential fix for pull request finding Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com> --------- Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
…1722) * agentHost: Preserve authority identity through URI serialization Encode non-readable remote addresses with lowercase hexadecimal so connection authorities remain stable across URI serialization and cannot collide across encoding tiers.\n\nFixes #331708.\n\n(Written by Copilot)\n\nCo-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * agentHost: Clarify authority encoding schema Document each encoding tier and make the reserved hex prefix check explicitly case-insensitive.\n\n(Written by Copilot)\n\nCo-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
* sessions: avoid transferring provisional sessions Only materialized sessions are shared when opening an editor window, keeping provisional Agents sessions scoped to their owning composer. Fixes #331592. (Written by Copilot) Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * sessions: document provisional handoff gate Correct the active-session type import and explain why only materialized sessions are shared across windows. (Written by Copilot) Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> --------- Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
* Chat input pills: reusable widget, artifacts, customizations and visibility Rework the pill row above the Agents-window chat input: - Left-align the row and make it horizontally scrollable, with a reusable observable-driven ChatPillsWidget in the workbench layer. Sessions owns the adapters from session state so the workbench layer never imports sessions. - Add chat.agentSessions.showSessionMetadataInInput, which moves the session header metadata pills down into the input row, hides the header second row, moves Chats into the title toolbar and shows workspace metadata inline. - Add agent-host artifact tools (add/remove/list_artifacts) with persistence, a gating setting and an artifacts pill. GitHub pull request and issue artifacts are promoted into the session GitHub links rather than shown twice. - Derive the customizations a chat used or read from its output stream and surface them in a customizations pill that reveals the picked entry in the customizations editor. - Add a right-click visibility menu for the row, with Hide <pill> for the clicked pill and kinds grouped by whether they have data. Customizations and Subagents start hidden; Changes can never be hidden. - Consolidate pill rendering onto one base plus four implementations: icon and label, dropdown, resource label, and the animated changes pill. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * also add this to the pr * Fix chat pill rendering crash from space-separated class names `classList.add` rejects tokens containing spaces, so the changes and resource pills threw while rendering. Subclasses now contribute a single modifier class instead of the full class list, which makes the mistake impossible, and the base always applies the shared classes. Port the changes pill's styling onto the shared pill classes and retire the now-dead chatTurnPills.css, whose rules all targeted the pre-refactor DOM. Cover the render path of every pill implementation, which is what CI caught and the existing unit tests missed. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Address PR review feedback on chat input pills - Restrict artifact `link` to http(s). A link is opened with `openExternal`, so a `file:` or custom-scheme link would reach the OS protocol handler from an agent-labelled pill. - Never let promotion lose an artifact. A GitHub reference is only removed from the artifacts pill when the GitHub pills actually surface it, so a session with no repository, or a reference belonging to another repository, keeps showing it. References from another repository are also no longer polled against the checkout's coordinates. - Let the dropdown pill's trigger close its own dropdown, and expose `aria-haspopup`/`aria-expanded` while it is summarized. - Keep the Windows drive prefix attached when matching customization paths, so `C:\repo\...` resolves. - Derive a plugin's container folder from its type rather than basename punctuation, so a versioned root such as `plugins/foo/1.2.0` no longer claims its sibling roots. - Gate customization data presence on the turn-status setting, without gating it on visibility, which would drop the pill from the menu that restores it. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Show session metadata above the chat input by default on insiders `chat.agentSessions.showSessionMetadataInInput` now defaults to on for non-stable builds, matching `chat.artifactTools.enabled`, which already uses the same gate. Stable keeps the session header's metadata row. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> --------- Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
* agentHost: preserve sessions during cold-start restore Prevent cold-start catalog migration from reporting valid sessions as missing. - Wait for the requested provider's initial catalog migration before session restore. - Keep tombstone checks ahead of migration and preserve failed-catalog retry behavior. - Keep the stored active chat until a delayed peer-chat catalog restores it. - Add regression tests for provider migration and delayed side-chat restoration. Fixes #331648 (Commit message generated by Copilot) Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * sessions: preserve pending chat during state save Keep the delayed active-chat identity when the window saves before the peer catalog arrives. - Store pending restored chat resources at the service level so snapshots preserve them. - Handle sessions whose active chat is temporarily unavailable during restoration. - Extend the regression test through another save and restart before catalog hydration. - Keep the concurrent-restore test focused on duplicated restore work. (Commit message generated by Copilot) Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> --------- Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…an MCP server (#331674) * Fix automation MCP launches against an installed build getApplication re-applied the from-source environment on every launch, so --build runs were also marked VSCODE_DEV=1. A packaged build then behaves as if it were running from a checkout and never opens a window, and the launch fails with Timeout 60000ms exceeded while waiting for event ''window''. That environment is already resolved at module load, and only for the from-source path, so drop the duplicate. The launcher also never passed an extensions directory, so a --build run loaded the user's installed extensions: those change the product under test and their logs are copied into the evidence bundle. Point it at an isolated directory under the existing test data path, matching the smoke test runner. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: adb443eb-11e5-40a1-8608-7f593fa79485 * Stop downloading a stable build the MCP server never uses \setup\ downloaded the previous stable release on every \--build\ launch, and the test data directory it downloads into is wiped at startup, so the cost was paid on every run: 335 MB and ~36s before VS Code even starts. The result was only ever assigned to \opts['stable-build']\, which nothing in this server reads. Migration tests belong to the smoke test runner, which has its own copy of this logic. Remove the download along with the now-unused \--stable-build\ option. A full \--build\ evidence run drops from about 90s to 19s. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: adb443eb-11e5-40a1-8608-7f593fa79485 * Run UI validation scenarios without configuring an MCP server The ui-scenario-validation skill could only be used by first registering test/mcp as an MCP server and then driving the evidence tools by hand. That is a lot of setup for a one-off reproduction, and it puts the capability out of reach of a plain request like ''reproduce this issue and record it''. Add runScenario, which takes a scenario file and runs it end to end: node test/mcp/out/runScenario.js <scenario.js> --build <app-root> A scenario is plain JavaScript that exports an id, a title and a list of steps. Each step gets the Application, the Workbench helpers and the Playwright page, returns a string describing how it was validated, throws to fail, or calls skip(reason) when a precondition is unavailable. The runner launches VS Code with video and tracing, captures a screenshot at every step boundary, stops at the first failed or skipped step, writes the report and renders the chapter titles onto the recording. Clean capture is enabled by default, so the recording shows unmodified UI and the step titles are added afterwards. A skipped step reports the run as aborted rather than passed: the scenario did not validate, so calling it a pass would overstate the result. Rewrite the skill around this flow. The MCP server remains available for interactive exploration, where inspecting the UI before choosing an assertion is worth the setup. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: adb443eb-11e5-40a1-8608-7f593fa79485 * Caption the recording instead of cutting chapter cards into it Step titles were shown on full-screen cards spliced between segments, which stretched the timeline and told the viewer nothing about what the step actually checked. The validation detail was meant to appear on those cards but never did: it was read from the opening capture, which only records that a step started, so the field was always empty. Draw a caption band under the frame instead. ffmpeg pads the canvas and the text is drawn into the added strip, so the recording keeps its original length and no recorded pixel is covered - the status bar and the bottom of the workbench stay readable. Each caption shows the step number and id, its status, the title, and the validation detail from the closing capture, coloured green, red or amber to match. The run outcome sits in the corner for the whole video. This also collapses the render to a single ffmpeg pass: the previous version trimmed one segment per step, generated a card per step and concatenated them. Give a CommonJS scenario a .cjs extension. This package is an ES module package, so a scenario saved as .js inside the repository is treated as ESM and a CommonJS one fails to load - which is exactly what the skill told people to do. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: adb443eb-11e5-40a1-8608-7f593fa79485 * Rename the skill to validate-ui-scenario Workspace skills are surfaced as slash commands named after the skill, so the name is what a user types. "ui-scenario-validation" reads as a topic; the verb form matches how it is invoked: /validate-ui-scenario reproduce <issue url> against my installed Insiders Update the example to that phrasing, and state that a reproduction which fails at the expected step is a successful reproduction rather than a broken scenario. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: adb443eb-11e5-40a1-8608-7f593fa79485 * Put the caption band above the recorded frame The band reads before the frame it describes, and it keeps the eye near the toolbar and editor where the action happens rather than at the far edge of the window. The canvas is still padded rather than overlaid, so no recorded pixel is hidden. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: adb443eb-11e5-40a1-8608-7f593fa79485 * Reject per-run arguments that escape the isolated profile Per-run `extraArgs` are appended after the generated arguments, and VS Code keeps the last value of a repeated string option, so a caller-supplied `--extensions-dir` would replace the isolated directory and let real user extensions and their logs back into the recording. The same applies to `--user-data-dir` on the restart path, which had no guard at all. Move the check into a shared helper and apply it to both the launch and restart argument paths. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: adb443eb-11e5-40a1-8608-7f593fa79485 * Unwrap a default export without the `in` operator `local/code-no-in-operator` rejects `'default' in loaded`. Reading the optional property directly is equivalent here and simpler: a CommonJS scenario has no `default`, and an ES module namespace carries the scenario on it. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: adb443eb-11e5-40a1-8608-7f593fa79485 --------- Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: adb443eb-11e5-40a1-8608-7f593fa79485
agentHost: restore deleted legacy session worktrees Restore the legacy metadata fallback that reconstructs the worktree path and repository root for pre-July 2026 Copilot sessions, allowing live sessions to recreate deleted worktrees on resume. (Written by Copilot) Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…sion (#331721) * agentHost: wait for provider catalog migration before restoring a session On Agents Window startup after an update, a restored session could show a persistent "Couldn't open session / Session not found on backend" error even though the session existed and loaded fine after navigating away and back. `AgentService.listSessions()` waits for each provider's registration-time catalog migration via `_awaitInitialProviderMigration()`, but `restoreSession()` did not. A restored editor subscribes during startup and triggers `restoreSession()` while the provider's initial catalog migration is still in flight; the provider's `getChatMetadata()` returns `undefined`, and `_restoreSessionState()` threw a false `AHP_SESSION_NOT_FOUND` that the client baked into synthetic chat history until re-subscription. Give restore the same gate: - Extract `_awaitInitialProviderMigrationForProvider(provider)` (reusing the existing compare-and-replace single-flight retry) and refactor the aggregate `_awaitInitialProviderMigration()` to delegate to it. - In `_doRestoreSession`, after the early tombstone check, await the session provider's initial migration, then re-check the tombstone before the registry/metadata work so a session deleted during the (potentially long) wait is not resurrected. - Classify the outcome truthfully: `_restoreSessionState` throws `AHP_SESSION_NOT_FOUND` only when the catalog was ready (a genuine miss), otherwise `JSON_RPC_INTERNAL_ERROR`, so a still-unavailable catalog is never reported as a missing session. Adds regression tests covering: waits-for-migration, delete-during-wait, genuinely-missing-after-migration, unavailable-catalog-is-internal-error, and provider specificity. Fixes #331648 Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * agentHost: classify restore metadata miss from an authoritative signal Address PR review feedback on #331721: - A metadata miss on restore is now treated as authoritative absence only when the provider catalog was readable this run AND the registry has no record of the session. A miss for a registered (known) session, or while the catalog was unavailable, is reported as a transient internal error rather than a sticky `AHP_SESSION_NOT_FOUND`. Previously `catalogReady` (derived solely from the one-time backfill promise) could be true on a backfilled restart while a provider was currently unavailable — e.g. Claude whose SDK is not downloaded yet returns `undefined` from getChatMetadata — and still emitted the sticky not-found. The registry is a download-free, authoritative record of existence, so no provider download is triggered. - Condense the method-body comments that exceeded the inline one-liner limit and the two test-double JSDocs. Adds a regression test for a registered session whose provider is currently unavailable on a backfilled restart (migration short-circuits): restore now rejects with an internal error, never a false not-found. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> --------- Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
* Removing managed settings check * updating types for sandbox configuration * fixing compile errors
* sticky scroll chat * potential sticky scroll fixes * add back rnderParsedRequestToPlainText * clamp height, editing value instead of input box * fix large gap + editing + address pr comments * address comments + fix css * fix tests * only show sticky scroll when fully hidden * address some more fixesgit add . * fix compiliation/hygiene * add fade in animation * Fix sticky scroll regressions Co-authored-by: justschen <54879025+justschen@users.noreply.github.com> * Fix sticky scroll review findings Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> --------- Co-authored-by: Rob Lourens <roblourens@gmail.com> Co-authored-by: BeniBenj <besimmonds@microsoft.com> Co-authored-by: Copilot <copilot@github.com> Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Defer idle Copilot configuration updates (Written by Copilot) Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
chat: avoid transferring local debug logs over IPC Keep local log and user-data resources as URI/size descriptors so the main process can stream them directly into the export. Bound unavoidable inline content for non-local resources to a shared 30 MiB budget.\n\n(Written by Copilot)\n\nCo-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
* make sure thinking headers have markdown rendering * Address thinking header review feedback Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Fix writable progress test fixture Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> --------- Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…he list (#331606) * agentHost: Make the first session listing authoritative for legacy Copilot CLI migration * agentHost: address PR review on legacy Copilot CLI migration readiness * agentHost: redirect legacy Copilot CLI opens through the agent host * agentHost: drop the startup discovery wait now that opens redirect * do not remember legacy adoption failures and cleanup * Feedback update * Few fixes and cleanup * test fix * Handle external session visibility and test fix * feedback updates * few updates
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
9 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
See Commits and Changes for more details.
Created by
pull[bot] (v2.0.0-alpha.4)
Can you help keep this open source service alive? 💖 Please sponsor : )