fix(tasks): rebind sandbox MCP session on actor transitions by VojtechBartos · Pull Request #70454 · PostHog/posthog

VojtechBartos · 2026-07-13T13:31:01Z

Problem

Even with every message carrying its sender (#70806, #70762), the sandbox's live MCP session doesn't follow speaker changes. Token refresh is gated by a per-run time window, so when user B speaks inside user A's window, B's turn runs on A's token and PostHog writes are attributed to (and authorized as) the wrong person. The window mark is also keyed by run id, so a replacement sandbox (retry, snapshot restore) inherits a stale mark.

Changes

One cache entry per sandbox — tasks:sandbox-mcp-session:{scope} — tracks which user the live session is bound to:

Same speaker within the window: skip (unchanged cost).
Speaker change, expired/unknown entry, or replacement sandbox: mint for the message's actor, push refresh_session, rebind. Unknown fails safe by refreshing.
Boot records the initial binding; nothing ever needs invalidating.

The gate reads the pinned actor from #70762's delivery, which the workflow serializes — so check-then-bind has a single executor.

Showcase

How did you test this code?

Tested manually: a two-user Slack thread where alternating speakers trigger a session refresh and same-speaker repeats skip it, with flag changes attributed to the correct user in PostHog. Automated gate tests cover same-actor skip, transition bypass, switch-back refresh, unknown-binding refresh, replacement-sandbox rebind, and binding only on successful delivery.

Automatic notifications

Publish to changelog?
Alert Sales and Marketing teams?

Docs update

Internal behavior only.

🤖 Agent context

Autonomy: Human-driven (agent-assisted)

Claude Code (Claude Fable 5).

trunk-io · 2026-07-13T13:47:08Z

_{View Full Report ↗︎ ⋅ Docs}

greptile-apps · 2026-07-14T09:54:42Z

Security Review

A no-config actor transition can be cached as if the sandbox was rebound, so later turns may continue using the previous actor's MCP session.

_{Reviews (1): Last reviewed commit: "fix(tasks): rebind sandbox MCP session o..." | Re-trigger Greptile}

github-actions · 2026-07-14T10:02:33Z

🤖 CI report

✅ Bundle size — no change

Uncompressed size of every built .js bundle, compared against the base branch.

Total: 64.72 MiB · no change

No file changed by more than 1000 B.

_{Posted automatically by build-bundle-size-report · uncompressed bytes from dist-report}

✅ Eager graph — within budget

How much code each root ships on the eager path — downloaded and parsed before the surface is interactive. Measured from the esbuild output chunks (post-tree-shake, static imports only); lazy import() / React.lazy chunks are not counted.

Root	Eager (shipped)	Δ vs base	Budget
entry (logged-out pages, app bootstrap) `src/index.tsx`	1.21 MiB · 22 files	no change	`███░░░░░░░` 28.2% of 4.29 MiB
authenticated shell (every logged-in page) `src/scenes/AuthenticatedShell.tsx`	8.12 MiB · 2,975 files	no change	`█████████░` 87.8% of 9.25 MiB

🟢 node_modules/monaco-editor/ stays out of src/index.tsx
🟢 src/lib/components/ActivityLog/describers stays out of src/index.tsx
🟢 [object Object] stays out of src/index.tsx
🟢 [object Object] stays out of src/index.tsx
🟢 node_modules/monaco-editor/ stays out of src/scenes/AuthenticatedShell.tsx
🟢 src/lib/components/ActivityLog/describers stays out of src/scenes/AuthenticatedShell.tsx
🟢 [object Object] stays out of src/scenes/AuthenticatedShell.tsx
🟢 [object Object] stays out of src/scenes/AuthenticatedShell.tsx

Largest files eagerly shipped from src/index.tsx

Size	File
126.8 KiB	`../node_modules/.pnpm/react-dom@18.3.1_react@18.3.1/node_modules/react-dom/cjs/react-dom.production.min.js`
24.6 KiB	`../node_modules/.pnpm/buffer@6.0.3/node_modules/buffer/index.js`
6.3 KiB	`../node_modules/.pnpm/react@18.3.1/node_modules/react/cjs/react.production.min.js`
4.5 KiB	`../node_modules/.pnpm/@jspm+core@2.1.0/node_modules/@jspm/core/nodelibs/browser/process.js`
3.9 KiB	`../node_modules/.pnpm/scheduler@0.23.2/node_modules/scheduler/cjs/scheduler.production.min.js`
1.4 KiB	`../node_modules/.pnpm/base64-js@1.5.1/node_modules/base64-js/index.js`
1.3 KiB	`src/RootErrorBoundary.tsx`
912 B	`../node_modules/.pnpm/ieee754@1.2.1/node_modules/ieee754/index.js`
789 B	`src/scenes/ChunkLoadErrorBoundary.tsx`
762 B	`src/index.tsx`

Largest files eagerly shipped from src/scenes/AuthenticatedShell.tsx

Size	File
280.3 KiB	`../node_modules/.pnpm/posthog-js@1.400.0/node_modules/posthog-js/dist/rrweb.js`
267.7 KiB	`../node_modules/.pnpm/@posthog+icons@0.38.0_react-dom@18.3.1_react@18.3.1__react@18.3.1/node_modules/@posthog/icons/dist/posthog-icons.es.js`
234.9 KiB	`src/taxonomy/core-filter-definitions-by-group.json`
222.8 KiB	`../node_modules/.pnpm/posthog-js@1.400.0/node_modules/posthog-js/dist/module.js`
164.0 KiB	`src/queries/validators.js`
154.3 KiB	`../node_modules/.pnpm/re2js@0.4.1/node_modules/re2js/build/index.esm.js`
126.8 KiB	`../node_modules/.pnpm/react-dom@18.3.1_react@18.3.1/node_modules/react-dom/cjs/react-dom.production.min.js`
105.9 KiB	`src/lib/api.ts`
93.3 KiB	`../node_modules/.pnpm/prosemirror-view@1.40.1/node_modules/prosemirror-view/dist/index.js`
90.6 KiB	`../node_modules/.pnpm/@tiptap+core@3.20.6_@tiptap+pm@3.20.6/node_modules/@tiptap/core/dist/index.js`

_{Posted automatically by check-eager-graph · sizes are eager output bytes (shipped, post-tree-shake) from the esbuild metafile · part of #32479}

✅ Dist folder size — no change

Total size of the built frontend/dist folder (all assets), compared against the base branch.

Total: 1284.95 MiB · no change

ℹ️ MCP UI apps size — 31 app(s), 16561.5 KB JS

Built size of each MCP UI app (main.js + styles.css).

App	JS	CSS
debug	598.2 KB	179.2 KB
action	456.5 KB	179.2 KB
action-list	563.0 KB	179.2 KB
cohort	455.4 KB	179.2 KB
cohort-list	562.0 KB	179.2 KB
email-template	455.3 KB	179.2 KB
error-details	471.1 KB	179.2 KB
error-issue	456.1 KB	179.2 KB
error-issue-list	562.9 KB	179.2 KB
experiment	560.1 KB	179.2 KB
experiment-list	563.8 KB	179.2 KB
experiment-results	561.8 KB	179.2 KB
feature-flag	565.8 KB	179.2 KB
feature-flag-list	569.5 KB	179.2 KB
feature-flag-testing	459.4 KB	179.2 KB
insight-actors	560.8 KB	179.2 KB
invite-email-preview	454.7 KB	179.2 KB
llm-costs	558.1 KB	179.2 KB
session-recording	457.2 KB	179.2 KB
session-summary	462.5 KB	179.2 KB
survey	457.0 KB	179.2 KB
survey-global-stats	560.9 KB	179.2 KB
survey-list	563.7 KB	179.2 KB
survey-stats	560.8 KB	179.2 KB
trace-span	455.8 KB	179.2 KB
trace-span-list	562.9 KB	179.2 KB
workflow	455.8 KB	179.2 KB
workflow-list	562.4 KB	179.2 KB
query-results	743.7 KB	179.2 KB
render-ui	824.0 KB	179.2 KB
visual-review-snapshots	460.3 KB	179.2 KB

⚠️

Backend coverage — 4.0% of changed backend lines covered — 119 uncovered

🧪 Backend test coverage

Patch coverage — changed backend lines (products + core): █░░░░░░░░░░░░░░░░░░░ 4.0% (5 / 124)

File	Patch	Uncovered changed lines
`products/tasks/backend/temporal/process_task/activities/send_followup_to_sandbox.py`	0.0%	166–168, 279, 281–288, 299, 324, 329, 332, 337, 339, 350, 352, 368, 370, 378
`products/tasks/backend/temporal/process_task/tests/test_send_followup_to_sandbox.py`	0.0%	27, 30, 32, 69–73, 76–80, 84–85, 139, 150, 152, 155, 163, 179, 186, 188, 199, 208, 214–218, 220–221, 223–224, 226, 233–236, 238–239, 241, 244–247, 250, 255, 258, 260, 265, 268, 270, 273, 275–276, 278–279, 282, 284, 286–287, 289, 292, 294, 296–297, 299, 302, 305, 307, 309–311, 313, 320, 322, 324–325, 327–329, 331, 338, 341–342, 344–346, 394
`products/tasks/backend/temporal/process_task/activities/start_agent_server.py`	33.3%	349–350
`products/tasks/backend/temporal/process_task/utils.py`	50.0%	371, 375, 384, 390

🤖 Agents: add a test covering the lines above, or note why under "How did you test this code?". Machine-readable gap list: the patch-coverage artifact on this run (gh run download 29739868576 -n patch-coverage), or the coverage-data block at the end of this comment.

Per-product line coverage (touched products)

Product	Coverage	Lines
`demo`	`███████████░░░░░░░░░` 56.2%	1,497 / 2,663
`tasks`	`█████████████░░░░░░░` 67.0%	25,953 / 38,752
`signals`	`████████████████░░░░` 79.4%	19,501 / 24,561
`data_modeling`	`████████████████░░░░` 80.0%	4,834 / 6,045
`cdp`	`████████████████░░░░` 80.7%	3,118 / 3,864
`notebooks`	`█████████████████░░░` 85.1%	7,096 / 8,338
`cohorts`	`█████████████████░░░` 86.2%	4,065 / 4,717
`agent_platform`	`█████████████████░░░` 86.4%	3,807 / 4,405
`actions`	`█████████████████░░░` 86.6%	717 / 828
`product_tours`	`█████████████████░░░` 87.5%	1,266 / 1,447
`exports`	`██████████████████░░` 88.4%	6,943 / 7,853
`conversations`	`██████████████████░░` 89.1%	16,361 / 18,360
`mcp_analytics`	`██████████████████░░` 89.2%	2,514 / 2,819
`dashboards`	`██████████████████░░` 89.4%	5,917 / 6,616
`engineering_analytics`	`██████████████████░░` 89.5%	5,507 / 6,154
`alerts`	`██████████████████░░` 89.9%	4,054 / 4,508
`error_tracking`	`██████████████████░░` 90.1%	9,757 / 10,828
`early_access_features`	`██████████████████░░` 90.1%	1,031 / 1,144
`streamlit_apps`	`██████████████████░░` 90.4%	2,501 / 2,767
`slack_app`	`██████████████████░░` 90.6%	8,989 / 9,926
`marketing_analytics`	`██████████████████░░` 90.8%	11,514 / 12,684
`stamphog`	`██████████████████░░` 91.0%	3,993 / 4,387
`product_analytics`	`██████████████████░░` 91.4%	5,781 / 6,328
`data_warehouse`	`██████████████████░░` 92.5%	18,819 / 20,348
`ai_observability`	`███████████████████░` 92.7%	14,921 / 16,093
`workflows`	`███████████████████░` 92.8%	5,520 / 5,949
`web_analytics`	`███████████████████░` 92.9%	13,853 / 14,913
`surveys`	`███████████████████░` 93.0%	5,724 / 6,157
`posthog_ai`	`███████████████████░` 93.2%	1,325 / 1,421
`approvals`	`███████████████████░` 93.3%	3,395 / 3,640
`reminders`	`███████████████████░` 93.4%	468 / 501
`endpoints`	`███████████████████░` 94.1%	8,606 / 9,143
`revenue_analytics`	`███████████████████░` 94.5%	3,598 / 3,809
`skills`	`███████████████████░` 94.5%	2,881 / 3,049
`review_hog`	`███████████████████░` 94.6%	6,806 / 7,191
`logs`	`███████████████████░` 95.3%	9,844 / 10,325
`experiments`	`███████████████████░` 95.7%	24,417 / 25,527
`replay_vision`	`███████████████████░` 95.8%	13,850 / 14,459
`annotations`	`███████████████████░` 96.2%	732 / 761
`feature_flags`	`███████████████████░` 96.3%	16,233 / 16,859
`user_interviews`	`███████████████████░` 96.4%	2,242 / 2,325
`warehouse_sources`	`███████████████████░` 96.5%	289,025 / 299,653
`customer_analytics`	`███████████████████░` 97.2%	7,482 / 7,700
`data_catalog`	`███████████████████░` 97.4%	2,303 / 2,365
`pulse`	`████████████████████` 98.4%	2,017 / 2,049

Report-only. Patch coverage = changed backend lines covered vs origin/master. Sorted lowest first.
Known gaps: lines covered only by Temporal tests show as uncovered; core line numbers may drift if master changed the same file.

✅ Hobby preview — passed

Hobby deployment smoke test passed successfully.

_{Run 29352414088}

VojtechBartos · 2026-07-14T15:17:28Z

This stack of pull requests is managed by Graphite. Learn more about stacking.

cvolzer3 · 2026-07-16T15:25:47Z

Non-blocking — one finding from reviewing the transition-with-no-configs path.

The comment at send_followup_to_sandbox.py ("an empty server list replaces it wholesale") doesn't match the agent-server's actual contract: the refresh_session handler short-circuits on an empty mcpServers list and returns { refreshed: true } without rebuilding the session, regardless of refreshedCredentials (agent-server.ts#L1180-L1182). The send_refresh_session docstring in this repo documents the same behavior, and the credentials-refresh loop (refresh_sandbox_credentials.py) depends on it — it sends [] every interval as a notification-only ping.

Consequence: on an actor transition where the new actor resolves zero configs, the empty-list send is a silent no-op that reports success, so mark_sandbox_mcp_session binds the new actor while the previous actor's session (including any user-level MCP servers) stays live — and the gate then skips every subsequent refresh for the new actor. test_transition_with_no_configs_still_clears_previous_session asserts what we send, not what the agent-server does, so it encodes the wrong assumption.

Why non-blocking: the branch is unreachable on Cloud — get_sandbox_ph_mcp_configs only returns empty when the MCP URL can't be resolved, which is a deployment-level condition, so wherever multi-user Slack threads run, mcp_configs is never empty. It's only reachable on deployments without a resolvable MCP URL.

Suggested follow-up:

On a transition that resolves no configs, don't send and don't mark — leave the previous actor's binding in place, since it accurately describes the live session (costs one wasted mint per message from that actor on affected deployments). Fix the comment and test to match.
If real teardown is wanted, it needs an explicit clear signal agent-server-side first — overloading empty-list would break the credentials-refresh loop — with sandbox-image rollout ordering respected before this repo can rely on it.

hex-security-app · 2026-07-20T09:17:48Z

+            # without a resolvable MCP URL (get_sandbox_ph_mcp_configs is
+            # otherwise never empty), so best-effort delivery is acceptable.
+            logger.info("refresh_mcp_no_configs_on_transition", run_id=run_id, previous_user_id=bound_user_id)
+            return True


Fail closed when a transition has no MCP configs

This path leaves the live session bound to previous_user_id but reports it safe, so _deliver_followup sends the new participant's prompt anyway. This is reachable when the PostHog MCP URL is unavailable and the new actor has no installations while the prior actor has a personal MCP installation: the agent can invoke that retained server with the prior actor's bearer token. Retaining the old cache binding prevents a later skip, but does not protect the current turn.

Prompt To Fix With AI

On an actor transition with an empty resolved MCP-config list, return False so _deliver_followup does not send a prompt against the retained prior-actor session. Preserve the previous binding and add a regression test that verifies the follow-up is rejected rather than delivered in this case.

_{Severity: medium | Confidence: 94% | React with 👍 if useful or 👎 if not}

The follow-up MCP refresh skipped whenever any token had been issued for the run within the freshness window, so when a different user spoke next (multiplayer Slack threads), the live session kept the previous actor's OAuth token for up to the rest of the window. One cache entry per sandbox (tasks:sandbox-mcp-session:{scope}) records which user the live session was last bound to and expires with the freshness window: same-actor repeats skip, while a speaker change, an expired or unknown entry, and a replacement sandbox (fresh scope) all refresh. The gate receives the delivery's already-resolved, payload- pinned actor, so it follows the message's sender; runs with no credential user skip the doomed mint instead of warning per message. Boot records the initial session binding under the sandbox id.

…fig transition An empty mcpServers list is a notification-only no-op on the agent-server (it returns without rebuilding the session), so sending it on an actor transition that resolves no configs neither clears the previous actor's session nor rebinds — yet it reported success and falsely marked the new actor. Stop sending it and leave the binding on the previous actor, who still holds the live session. Only reachable on deployments without a resolvable MCP URL.

The session marker self-expires at half the OAuth token lifetime, so an absent marker can mean the previous actor's session is still live, not that the sandbox is fresh. Treating that as a best-effort first bind delivered a new actor's follow-up under the previous actor's credentials when the mint or refresh failed in the marker-expired window. Fail closed whenever a rebind can't be confirmed, not only on a known transition.

graphite-app · 2026-07-20T10:36:20Z

    def test_skips_send_when_no_mcp_configs_resolved(
-        self, mock_oauth, mock_ph_configs, mock_user_configs, mock_send_refresh
+        self, mock_oauth, mock_ph_configs, mock_user_configs, mock_send_refresh, _sleep
    ):
-        mock_oauth.return_value = "fresh-token"
+        _arm_success(mock_oauth, mock_ph_configs, mock_user_configs, mock_send_refresh)
        mock_ph_configs.return_value = []
-        mock_user_configs.return_value = []

        _refresh(_make_task_run_mock())

        mock_send_refresh.assert_not_called()
+        # Marked anyway: with no session to rebind, don't re-mint per message.
+        assert get_sandbox_mcp_session_user("run-1") == 42



Test assertion doesn't match production code behavior. When mcp_configs is empty, the production code at line 324-334 of send_followup_to_sandbox.py returns False without calling mark_sandbox_mcp_session(). Therefore, the cache will remain empty and the assertion assert get_sandbox_mcp_session_user("run-1") == 42 will fail because it expects the value to be 42 but will get None.

# The assertion should be: assert get_sandbox_mcp_session_user("run-1") is None

Or if the intent is to mark even with no configs, the production code needs to call mark_sandbox_mcp_session(scope, actor_user.id) before returning at line 334.

Suggested change

def test_skips_send_when_no_mcp_configs_resolved(

self, mock_oauth, mock_ph_configs, mock_user_configs, mock_send_refresh

self, mock_oauth, mock_ph_configs, mock_user_configs, mock_send_refresh, _sleep

):

mock_oauth.return_value = "fresh-token"

_arm_success(mock_oauth, mock_ph_configs, mock_user_configs, mock_send_refresh)

mock_ph_configs.return_value = []

mock_user_configs.return_value = []

_refresh(_make_task_run_mock())

mock_send_refresh.assert_not_called()

# Marked anyway: with no session to rebind, don't re-mint per message.

assert get_sandbox_mcp_session_user("run-1") == 42

def test_skips_send_when_no_mcp_configs_resolved(

self, mock_oauth, mock_ph_configs, mock_user_configs, mock_send_refresh, _sleep

):

_arm_success(mock_oauth, mock_ph_configs, mock_user_configs, mock_send_refresh)

mock_ph_configs.return_value = []

_refresh(_make_task_run_mock())

mock_send_refresh.assert_not_called()

# Not marked: with no configs resolved, mark_sandbox_mcp_session is not called.

assert get_sandbox_mcp_session_user("run-1") is None

Spotted by Graphite

Is this helpful? React 👍 or 👎 to let us know.

On an actor transition where the new actor resolves no MCP configs, an empty-list refresh_session is a no-op on the agent-server: it can neither rebind the sandbox MCP session to the new actor nor tear down the previous actor's. Returning True let _deliver_followup run the new actor's turn against a session that may still hold the prior actor's credentials, so fail closed. A first/unknown binding with no configs has no recorded prior actor and nothing to leak, so it still runs the turn — the agent is not blocked just because MCP is unavailable (only self-hosted deployments without a resolvable MCP URL reach an empty config list at all).

hex-security-app · 2026-07-20T10:41:25Z

+        # there is nothing to leak, so let the turn run rather than block the
+        # agent just because MCP is unavailable. Record the binding so a later
+        # actor transition is still detected.
+        mark_sandbox_mcp_session(scope, actor_user.id)


Fail closed for unknown bindings with no configs

None is not evidence that this sandbox has no live MCP session: this marker expires after three hours while its OAuth token is valid for six. After actor A's marker expires, actor B on a deployment without a resolvable PostHog MCP URL can resolve no configs, take this branch, be marked as the session owner without a refresh_session, and have their prompt delivered using A's retained personal MCP bearer token. Subsequent B turns then skip rebinding for another cache window.

Prompt To Fix With AI

Treat an empty resolved MCP-config list as unsafe when the session binding is unknown as well as when it belongs to another actor. Do not update the session-binding cache and return False so `_deliver_followup` rejects the turn, unless durable sandbox state proves no MCP session was ever initialized. Add a regression test for an expired/evicted marker with a retained prior-actor session and no configs for the new actor.

_{Severity: medium | Confidence: 98% | React with 👍 if useful or 👎 if not}

_refresh_sandbox_mcp is now a bool-returning gate: _deliver_followup rejects the turn when it returns falsy. The refresh-ordering test stubbed it with a side effect that returned None, which trips the gate. Return True so the stub honors the contract and the test exercises ordering, not the reject path.

deployment-status-posthog · 2026-07-20T12:37:04Z

Deploy status

Environment	Status	Deployed At	Workflow
dev	✅ Deployed	2026-07-20 12:37 UTC	Run
prod-us	✅ Deployed	2026-07-20 12:48 UTC	Run
prod-eu	✅ Deployed	2026-07-20 12:53 UTC	Run

VojtechBartos self-assigned this Jul 13, 2026

This was referenced Jul 14, 2026

feat(slack-app): bind sandbox MCP identity to each message's actor #69618

Closed

feat(tasks): rebind sandbox GitHub identity and git author to the actor #70024

Closed

VojtechBartos marked this pull request as ready for review July 14, 2026 09:52

pr-assigner-resolver-posthog Bot requested a review from a team July 14, 2026 09:52

greptile-apps Bot reviewed Jul 14, 2026

View reviewed changes

Comment thread products/tasks/backend/temporal/process_task/activities/send_followup_to_sandbox.py Outdated

VojtechBartos marked this pull request as draft July 14, 2026 09:55

VojtechBartos mentioned this pull request Jul 14, 2026

feat(tasks): rebind sandbox GitHub identity on actor transitions #70668

Closed

2 tasks

VojtechBartos force-pushed the vojtab/mcp-session-identity-transitions branch from 4969cdc to 8da27f6 Compare July 14, 2026 15:15

VojtechBartos changed the base branch from master to graphite-base/70454 July 14, 2026 15:16

VojtechBartos force-pushed the graphite-base/70454 branch from ab8d961 to 5df4389 Compare July 14, 2026 15:17

VojtechBartos changed the base branch from graphite-base/70454 to vojtab/slack-followup-queue-actor July 14, 2026 15:17

VojtechBartos mentioned this pull request Jul 14, 2026

feat(tasks): route slack follow-ups through the queue with per-message actors #70762

Merged

2 tasks

VojtechBartos force-pushed the vojtab/mcp-session-identity-transitions branch 2 times, most recently from 7e00cc6 to 8d0d872 Compare July 14, 2026 15:47

VojtechBartos force-pushed the vojtab/slack-followup-queue-actor branch from e307e5b to 6558def Compare July 14, 2026 17:04

VojtechBartos force-pushed the vojtab/mcp-session-identity-transitions branch from 8d0d872 to f2b815c Compare July 14, 2026 17:06

VojtechBartos changed the base branch from vojtab/slack-followup-queue-actor to graphite-base/70454 July 14, 2026 17:06

VojtechBartos mentioned this pull request Jul 14, 2026

feat(tasks): route slack follow-ups through the workflow follow-up queue #70806

Merged

2 tasks

VojtechBartos force-pushed the vojtab/mcp-session-identity-transitions branch 7 times, most recently from 540c261 to 1bd42a4 Compare July 16, 2026 12:07

VojtechBartos mentioned this pull request Jul 16, 2026

feat(tasks): widen send_followup_message signal handler arity #71562

Merged

2 tasks

cvolzer3 approved these changes Jul 16, 2026

View reviewed changes

VojtechBartos force-pushed the vojtab/slack-followup-queue-actor branch from f7baa33 to a8a59c9 Compare July 17, 2026 09:38

VojtechBartos force-pushed the vojtab/mcp-session-identity-transitions branch from 8bbd3c9 to 5b41a75 Compare July 17, 2026 09:38

VojtechBartos force-pushed the vojtab/slack-followup-queue-actor branch from a8a59c9 to d9caad1 Compare July 17, 2026 10:08

VojtechBartos force-pushed the vojtab/mcp-session-identity-transitions branch 2 times, most recently from 230c58f to 4a56c2a Compare July 17, 2026 11:42

veria-ai Bot reviewed Jul 17, 2026

View reviewed changes

Comment thread products/tasks/backend/temporal/process_task/activities/send_followup_to_sandbox.py

VojtechBartos force-pushed the vojtab/slack-followup-queue-actor branch from d9caad1 to 0c1ff7e Compare July 20, 2026 07:48

VojtechBartos force-pushed the vojtab/mcp-session-identity-transitions branch 2 times, most recently from 8b6e210 to 92badb0 Compare July 20, 2026 08:45

VojtechBartos force-pushed the vojtab/slack-followup-queue-actor branch from 0c1ff7e to 632c21f Compare July 20, 2026 08:45

VojtechBartos force-pushed the vojtab/mcp-session-identity-transitions branch from 92badb0 to feec41d Compare July 20, 2026 09:16

hex-security-app Bot reviewed Jul 20, 2026

View reviewed changes

VojtechBartos force-pushed the vojtab/mcp-session-identity-transitions branch from feec41d to 54a3240 Compare July 20, 2026 09:21

Base automatically changed from vojtab/slack-followup-queue-actor to master July 20, 2026 10:10

VojtechBartos added 3 commits July 20, 2026 12:11

VojtechBartos force-pushed the vojtab/mcp-session-identity-transitions branch from 54a3240 to 02268db Compare July 20, 2026 10:12

hex-security-app Bot reviewed Jul 20, 2026

View reviewed changes

Comment thread products/tasks/backend/temporal/process_task/activities/send_followup_to_sandbox.py

graphite-app Bot reviewed Jul 20, 2026

View reviewed changes

VojtechBartos force-pushed the vojtab/mcp-session-identity-transitions branch from 454d497 to 0294158 Compare July 20, 2026 10:40

hex-security-app Bot reviewed Jul 20, 2026

View reviewed changes

VojtechBartos merged commit 116839d into master Jul 20, 2026
234 checks passed

VojtechBartos deleted the vojtab/mcp-session-identity-transitions branch July 20, 2026 12:14

Navigation Menu

Uh oh!

fix(tasks): rebind sandbox MCP session on actor transitions - #70454

fix(tasks): rebind sandbox MCP session on actor transitions#70454
VojtechBartos merged 5 commits into
masterfrom
vojtab/mcp-session-identity-transitions

VojtechBartos commented Jul 13, 2026 •

edited

Loading

Uh oh!

trunk-io Bot commented Jul 13, 2026 •

edited

Loading

Uh oh!

greptile-apps Bot commented Jul 14, 2026

Security Review

Uh oh!

Uh oh!

github-actions Bot commented Jul 14, 2026 •

edited

Loading

🧪 Backend test coverage

Uh oh!

VojtechBartos commented Jul 14, 2026 •

edited

Loading

Uh oh!

cvolzer3 commented Jul 16, 2026

Uh oh!

Uh oh!

hex-security-app Bot Jul 20, 2026

Uh oh!

Uh oh!

graphite-app Bot Jul 20, 2026

Uh oh!

hex-security-app Bot Jul 20, 2026

Uh oh!

Uh oh!

deployment-status-posthog Bot commented Jul 20, 2026 •

edited

Loading

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

VojtechBartos commented Jul 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Problem

Changes

Showcase

How did you test this code?

Automatic notifications

Docs update

🤖 Agent context

Uh oh!

trunk-io Bot commented Jul 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

greptile-apps Bot commented Jul 14, 2026

Security Review

Uh oh!

Uh oh!

github-actions Bot commented Jul 14, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🤖 CI report

🧪 Backend test coverage

Uh oh!

VojtechBartos commented Jul 14, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

cvolzer3 commented Jul 16, 2026

Uh oh!

Uh oh!

hex-security-app Bot Jul 20, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

graphite-app Bot Jul 20, 2026

Choose a reason for hiding this comment

Uh oh!

hex-security-app Bot Jul 20, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

deployment-status-posthog Bot commented Jul 20, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Deploy status

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

VojtechBartos commented Jul 13, 2026 •

edited

Loading

trunk-io Bot commented Jul 13, 2026 •

edited

Loading

github-actions Bot commented Jul 14, 2026 •

edited

Loading

VojtechBartos commented Jul 14, 2026 •

edited

Loading

deployment-status-posthog Bot commented Jul 20, 2026 •

edited

Loading