QUALITY-726: session sharing for orchestrated agent sessions by cephalonaut · Pull Request #11465 · warpdotdev/warp

cephalonaut · 2026-05-21T04:15:23Z

What

Makes the orchestration pill bar work for shared-session viewers across every orchestrator/child topology (local-local, local-remote, remote-local, remote-remote × share-parent / share-child). See specs/QUALITY-726/PRODUCT.md and specs/QUALITY-726/TECH.md for the full design.

User-visible changes:

A manually shared local orchestrator now drives orchestration discovery in its viewer's pill bar (previously the gate matched only AmbientAgent, so manually-shared local orchestrators got no pill bar).
Local children of a manually-shared local orchestrator inherit the share automatically, and the share's children pane shows them. Children that existed before the share started are caught up via a one-shot cascade. Stopping the host's share cascades to those children.
Local-to-driver children of a remote orchestrator now surface in the parent's pill bar — the local client reports the freshly-minted session_id back to warp-server on share so the viewer's REST descendant fetch can locate it.
Agent-id references inside conversation bodies resolve to the correct display name on first poll (previously the viewer-side index was never populated, so references rendered as "Unknown").
Non-orchestrator user shares stop polling after their first empty descendant fetch and only resume on the next AppendedExchange on the orchestrator.

Wire-format notes:

SessionSourceType::User stays a strict unit variant on the wire. A new source_task_id: Option<String> sidecar travels alongside source_type on InitPayload and JoinedSuccessfully (and the server's SessionManifest), carrying the conversation's ai_tasks row id when the user shares an orchestrator. This avoids the wire-incompatibility an earlier iteration introduced when it turned User into a struct variant.
A new BlocklistAIHistoryEvent::LocalSharedSessionEstablished lifecycle event drives a sidecar model that links the local share's session id to its ai_tasks row via update_agent_task.

Demos

local → local: https://www.loom.com/share/7cab1a115706404c9869a41d28dc1efd
local → remote: https://www.loom.com/share/e06c7d4da56f4923b9304f94e4982a29
remote → local: https://www.loom.com/share/57ae334e3d5f494393709ac87606755b
remote → remote: https://www.loom.com/share/b66e778549a14108be6d9d440be1b3cf

Known issues / follow-up

Child-link sibling resolution deferred. Opening a child agent's share link directly does not yet resolve sibling agent IDs in the child transcript — they fall back to the existing "Unknown agent" placeholder. The original design fetched siblings on join and registered them via start_new_conversation, which polluted live_conversation_ids_for_terminal_view, emitted StartedNewConversation events that cascaded into session-restoration persistence, and exposed siblings as navigable conversations in pickers. Stripped from this PR; a follow-up should add a name-only resolution path on BlocklistAIHistoryModel that doesn't go through start_new_conversation.
Pre-StreamInit share with concurrent viewer join. If the user shares a brand-new local conversation before the first response (so task_id is still None) and a viewer joins in that window, the sharer's later StreamInit upgrade mutates only the local TerminalModel.shared_session_source.source_task_id. Already-joined viewers received source_task_id: None on JoinedSuccessfully and have no push channel for the upgraded value, so their OrchestrationViewerModel never constructs and they never see the pill bar for that share. Workaround: the affected viewer can rejoin to pick up the upgraded value. Follow-up: add a sharer→viewer DownstreamMessage::UpdateSourceTaskId (or piggyback on an existing channel) and re-run viewer-side orchestration-discovery construction on receipt. Narrow blast radius — common-case shares and late-joining viewers are unaffected.
Catch-up cascade dispatch test coverage. The inherit_share_for_local_child rule itself has tests; the pane-group dispatch loop does not yet.

Related PRs (must land together)

session-sharing-protocol wire-format change: Add source_task_id sidecar; keep User as a strict unit variant session-sharing-protocol#15
session-sharing-server vendored-protocol mirror: https://github.com/warpdotdev/session-sharing-server/pull/449
warp-server session-link gate relaxation: https://github.com/warpdotdev/warp-server/pull/11372

The Cargo.toml [patch] override in this PR currently points the protocol crate at a local sibling worktree; remove the override and bump the rev pin once warpdotdev/session-sharing-protocol#15 lands on main.

Agent Mode

This PR was created by Warp Agent Mode

Co-Authored-By: Oz oz-agent@warp.dev

oz-for-oss · 2026-05-21T12:12:29Z

@cephalonaut

I'm starting a first review of this pull request.

You can view the conversation on Warp.

I completed the review and no human review was requested for this pull request.

Comment /oz-review on this pull request to retrigger a review (up to 3 times on the same pull request).

Powered by Oz

oz-for-oss

Overview

This PR extends shared-session orchestration discovery across local/remote parent and child topologies, adds server-side session-id linking for local shared sessions, and adds specs/tests for the new behavior.

Concerns

The committed Cargo.toml local path patch makes the branch unreproducible outside a sibling checkout and overrides the pinned protocol revision.
A manual share that starts before the conversation task id exists only upgrades the local model source type; existing viewers remain on User { task_id: None } and never start orchestration discovery.
Child-link sibling preload creates normal live conversations for siblings, which can expose sibling navigation/metadata in a view that is supposed to stay child-scoped.

Verdict

Found: 0 critical, 3 important, 0 suggestions

Request changes

Comment /oz-review on this pull request to retrigger a review (up to 3 times on the same pull request).

Powered by Oz

oz-for-oss · 2026-05-21T12:16:28Z

+# on `main`; patch to the sibling local worktree until the upstream PR lands.
+# Remove this table once that happens.
+[patch."https://github.com/warpdotdev/session-sharing-protocol"]
+session-sharing-protocol = { path = "../session-sharing-protocol" }


⚠️ [IMPORTANT] This local path patch makes the branch unbuildable outside a checkout that has ../session-sharing-protocol, and it overrides the pinned git rev. Remove the patch and commit the upstream protocol rev once the protocol PR lands before merging.

Will fix when protocol PR lands

oz-for-oss · 2026-05-21T12:16:28Z

+                            return;
+                        };
+
+                        model


⚠️ [IMPORTANT] Updating only the local TerminalModel leaves already-joined viewers with User { task_id: None }, so a share opened before StreamInit never constructs OrchestrationViewerModel after the task id arrives. Propagate this source-type upgrade to active viewers or avoid advertising the pre-StreamInit path as parent-scoped until reconnect.

Narrow blast radius - will follow up.

oz-for-oss · 2026-05-21T12:16:28Z

+                // Standalone conversation: there's no local parent on a
+                // child-link viewer, so we don't try to wire one up. The
+                // conversation is purely a name-resolution placeholder.
+                let conversation_id = history.start_new_conversation(


⚠️ [IMPORTANT] start_new_conversation registers each sibling in live_conversation_ids_for_terminal_view and emits StartedNewConversation, so a child-link viewer can gain real sibling conversations even though child links must stay scoped to the joined child. Populate the agent-id/name index with metadata-only placeholders or otherwise keep these sibling entries non-navigable.

fixed - by removing the sibling related code for now

## What Adds a `source_task_id: Option<String>` sidecar field to `sharer::InitPayload` and `viewer::DownstreamMessage::JoinedSuccessfully`. `SessionSourceType::User` stays a strict unit variant (no struct fields), matching `main`. `AmbientAgent { task_id }` is unchanged. Both new fields are gated on `#[serde(default)]` so older clients and viewers ignore them silently. The sidecar carries the conversation's server-side `ai_tasks` row id so downstream orchestration discovery can enumerate descendants without re-shaping the variant. ## Why Required by the warp client to support session sharing of *manually-shared* local orchestrator conversations. ## Related - session-sharing-server vendored-protocol mirror: warpdotdev/session-sharing-server#449 - warp client: warpdotdev/warp#11465 Co-Authored-By: Oz <oz-agent@warp.dev> --------- Co-authored-by: Oz <oz-agent@warp.dev>

advait-m

nice, the cascade approach makes sense - cool to see it working across all the permutations!

advait-m · 2026-05-21T21:11:19Z

+            });
+        }
+
+        // Suppress unused-import warning on wasm builds where the helper


nit: does this mean non-wasm? though, in general, not sure if I see why this or the use crate::ai::blocklist::orchestration_topology; import is needed.

advait-m · 2026-05-21T21:13:23Z

+            return;
+        };
+        let Some(host_source) =
+            crate::pane_group::pane::terminal_pane::host_terminal_shared_session_source_type(


nit: move imports to top?

advait-m · 2026-05-21T21:13:51Z

+
+        // Emit for a conversation that was never registered: the subscriber
+        // must early-return without firing an RPC.
+        let bogus_conversation_id = crate::ai::agent::conversation::AIConversationId::new();


nit: move import to top?

advait-m · 2026-05-21T21:31:22Z

+## Parallelization
+The four threads touch largely independent code areas. They run as parallel local sub-agents with separate worktrees once Thread A0 lands, then merge into a single PR (or 1-PR-per-thread). The strict sequencing is: A0 (protocol crate) → A/B/C/D in parallel.


This was cool to read - it sounds like the parallelization worked well in practice? @cephalonaut

I think it went pretty well! Would definitely be interested in feedback if you try it yourself.

nice; will let you know!

Address Oz auto-review concern #1 on PR #11465: the temporary `[patch."https://github.com/warpdotdev/session-sharing-protocol"]` override pointing at `../session-sharing-protocol` made the branch unbuildable outside a sibling checkout and shadowed the pinned git rev. QUALITY-726 has now landed on session-sharing-protocol `main` as 5a0ad6135809feee9da2e9efae8bd6b54b89172e (PR #15). This commit: - Bumps the workspace `session-sharing-protocol` dep to that rev. - Removes both `[patch]` tables that overrode the upstream URL with the local sibling worktree. Co-Authored-By: Oz <oz-agent@warp.dev>

…al and remote topologies Net effect of the QUALITY-726 integration branch, squashed. - Threads a `source_task_id` orchestrator-task-id sidecar through the shared-session protocol (`InitPayload`, `JoinedSuccessfully`, `SessionManifest`) and through the warp client so: - Manually-shared local orchestrator panes carry a task id when one is available, enabling orchestration discovery for their viewers. - Local-spawned children inherit the host's share with the child's task id stamped on the sidecar. - Bundles `source_type` + `source_task_id` into a single `SharedSessionSource` carried on `IsSharedSessionCreator::Yes`, `SharedSessionStatus::SharePendingPreBootstrap`, and `TerminalModel`. - Adds a catch-up cascade so child agent panes that existed before the parent started sharing also get auto-shared, via `BlocklistAIHistoryEvent::LocalSharedSessionEstablished` and `transitively_share_existing_local_children`. Stops auto-shared children when the host stops sharing. - Adds a driver-side session-id link: `LocalSharedSessionLinkModel` pings `update_agent_task` with the new `session_id` for local orchestrator conversations so the warp-server can surface the joinable link. - Drops a per-conversation dedupe `HashSet` from `LocalSharedSessionLinkModel` (server is now the source of truth). - Pins `session-sharing-protocol` to the merged main rev 5a0ad61 (PR #15) — removed the temporary `[patch]` override to the sibling worktree. - Hoists `host_terminal_shared_session_source_type` and `inherit_share_for_local_child` to a top-of-file cfg-gated `use` in `pane_group/mod.rs` and removes a dead `orchestration_topology` import + suppression line. Adds `specs/QUALITY-726/PRODUCT.md` and `TECH.md` documenting the design and parallelization approach (Threads A/B/C/D ran as parallel sub-agents after the protocol crate landed). - Pre-StreamInit viewer-join edge case: `source_type` upgrade on the host only mutates the local `TerminalModel`; viewers that joined before `StreamInit` stay on `User { task_id: None }` until reconnect. Tracked as a follow-up. - `updateAgentConversationIDQuery` still gates strictly on `state=RUNNING` server-side; same-class latent bug as the session-link case. To be addressed by the planned follow-up that starts local executions in RUNNING. Co-Authored-By: Oz <oz-agent@warp.dev>

When the sharer builds historical scrollback for a late-joining viewer, `reconstruct_response_events_from_conversations` was unconditionally emitting a Finished event after every exchange, including ones still in-flight. The viewer would take its `current_response_id` slot on the replayed Finished and then drop every live ClientAction arriving for the same still-running stream — the conversation appeared stuck on the initial prompt until the next request started. Gate the synthetic Finished on `exchange.output_status.is_finished()` so in-flight exchanges end the scrollback with their accumulated messages only; the live wire's real Finished closes the stream naturally. Co-Authored-By: Oz <oz-agent@warp.dev>

`BlocklistAIHistoryEvent` (pane_group/mod.rs) and `SharedSessionSource` (pane_group/pane/terminal_pane.rs) are only consumed inside `#[cfg(not(target_family = "wasm"))]` blocks, so the wasm build flags them as unused. `PaneGroup::stop_transitively_shared_child_shares` is only invoked from a non-wasm dispatch arm, so the wasm build flags it as dead. Gate each with the matching cfg attribute to keep the wasm clippy build warning-clean. Co-Authored-By: Oz <oz-agent@warp.dev>

The rebase ran the workspace's default rustfmt against several files whose import blocks then drifted from master's organization, producing 800+ lines of import-only churn in the PR diff. Re-formatting these four files with rustfmt's stable form (`imports_granularity=Module`, `group_imports=StdExternalCrate`) restores their original master-side import layout. No semantic changes — diff vs master shrinks by ~800 lines on these four files alone: app/src/terminal/view.rs 863 → 38 lines app/src/terminal/model/terminal_model.rs 166 → 46 lines app/src/terminal/local_tty/terminal_manager.rs 238 → 119 lines app/src/workspace/view_tests.rs 101 → 13 lines Co-Authored-By: Oz <oz-agent@warp.dev>

cla-bot Bot added the cla-signed label May 21, 2026

cephalonaut mentioned this pull request May 21, 2026

Add source_task_id sidecar; keep User as a strict unit variant warpdotdev/session-sharing-protocol#15

Merged

cephalonaut marked this pull request as ready for review May 21, 2026 12:12

oz-for-oss Bot reviewed May 21, 2026

View reviewed changes

cephalonaut requested review from advait-m and szgupta and removed request for szgupta May 21, 2026 16:50

advait-m approved these changes May 21, 2026

View reviewed changes

cephalonaut force-pushed the matthew/QUALITY-726-integration branch from 0136abb to 83ea458 Compare May 22, 2026 02:15

cephalonaut and others added 2 commits May 22, 2026 03:21

cephalonaut force-pushed the matthew/QUALITY-726-integration branch from 97d5d3f to e23b5e9 Compare May 22, 2026 07:47

cephalonaut enabled auto-merge (squash) May 22, 2026 07:56

cephalonaut merged commit 81d3496 into master May 22, 2026
25 checks passed

cephalonaut deleted the matthew/QUALITY-726-integration branch May 22, 2026 08:10

+                                          return;
+                                      };
+                                      model

		## Parallelization
		The four threads touch largely independent code areas. They run as parallel local sub-agents with separate worktrees once Thread A0 lands, then merge into a single PR (or 1-PR-per-thread). The strict sequencing is: A0 (protocol crate) → A/B/C/D in parallel.

Conversation

cephalonaut commented May 21, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What

Demos

Known issues / follow-up

Related PRs (must land together)

Agent Mode

Uh oh!

oz-for-oss Bot commented May 21, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

oz-for-oss Bot left a comment

Choose a reason for hiding this comment

Overview

Concerns

Verdict

Uh oh!

oz-for-oss Bot May 21, 2026

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

oz-for-oss Bot May 21, 2026

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

oz-for-oss Bot May 21, 2026

Choose a reason for hiding this comment

Uh oh!

cephalonaut May 21, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

advait-m left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

cephalonaut commented May 21, 2026 •

edited

Loading

oz-for-oss Bot commented May 21, 2026 •

edited

Loading

cephalonaut May 21, 2026 •

edited

Loading