[dustinhealy]

Summary

Agent sharing permission currently lets a member claim an unassigned destination for any organization-installed Slack bot. A connection's creator then controls replacement, rather than an administrator choosing who may manage the bot.

Administrators can now select Manage bot access on a connected workspace and grant named users management of that app/workspace installation. Those managers can assign agents and explicitly replace another manager's assignment, provided they also have permission to share the agent they choose. Managing a bot does not grant Slack credentials, installation administration, Agent access or permission to manage other managers.

How it works

• Reuses tenant-scoped ACL entries with a distinct slackBot resource and direct user grants. Dedicated, fresh-admin-authorized endpoints manage membership; generic sharing does not support this resource.
• Checks bot management inside the existing assignment transaction. Grant updates and assignments write the same workspace authority fence, so an assignment paused before revocation cannot commit afterward. Revision-checked updates prevent stale dialogs from overwriting another administrator's changes.
• Offers members only bots they can manage, filtering before pagination. Manager search uses the same fresh administrator capabilities as grant updates, the existing user-search helpers and a tenant/account-scoped cache. It returns only current users from the active tenant.
• Preserves existing assignments and conversations after management access is removed. A former manager can still remove assignments or delete their own connection. Active reauthorization preserves grants; disconnect and fresh installation clear them.

Administrator grants bot management
  → manager chooses a permitted bot in Agent Builder
  → Agent SHARE and Remote Agents USE remain required
  → assignment transaction checks the current bot grant
  → Slack senders still link and use their own LibreChat accounts

Testing

• Real MongoDB tests cover tenant-local grants, stale revisions, paused assignment versus revocation, co-manager replacement, filtered pagination, token refresh and disconnect/reinstallation.
• HTTP tests cover capability-authorized administrators without legacy VIEW_USERS, tenant-local candidate filtering, administrator-only grants, forged permissions, member assignment and cleanup after revocation; generic sharing remains denied.
• Component tests cover selecting managers, cancellation, stale-save recovery, failed reads, visible search failure/retry and hiding ungranted bots.
• Chrome exercises installation, manager selection/save/reopen, member assignment/replacement and cleanup in light and dark mode. External Slack/workspace HTTP responses are fixtures; storage authorization is exercised separately against MongoDB.
• Provider configuration tests, owning workspace builds/typechecks, full pre-push checks and Lighthouse passed locally. The hook skips config-migration tests and optional depcheck is unavailable; the provider configuration suite passed all 601 cases.

Screenshots / recordings

The bot-access dialog is new; there is no corresponding dialog on the base revision. These captures show the running application with a selected manager, before saving.

Grant bot management

Light	Dark

Retry a failed search

A failed lookup displays an actionable error and keeps unsaved management changes in the dialog.

Light	Dark

Risk / compatibility

Non-administrators require an explicit bot-manager grant before creating or changing assignments, including previously unclaimed destinations. Existing assignments keep running. The configurable channels.maxBotManagers defaults to 50 per installation. No deployment configuration changes are required.

Checklist

• I reviewed my own changes
• Relevant tests have been added or updated
• Existing relevant tests pass
• The change does not introduce new warnings or errors
• User-facing or complex behavior is documented where necessary
• Required dependency changes have been merged/published
• Required documentation PR: LibreChat documentation #892

[dustinhealy]

@codex review

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-11-05T16:42:38.483496Z	2d87f24	Manual request

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 19fce38ac8

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[github-actions]

Lighthouse CI failed. The last 80 log lines contain the measured budgets and assertion failures.

│ 23      │ 'http://localhost:3080/api/mcp/servers'                                                                         │ 2971.95000000007   │ 4239.038000000059  │ 200    │
│ 24      │ 'http://localhost:3080/api/permissions/mcpServer/effective/all'                                                 │ 2973.094000000041  │ 3739.6040000000503 │ 200    │
│ 25      │ 'http://localhost:3080/api/prompts/groups?limit=10'                                                             │ 2973.286000000022  │ 4244.257000000041  │ 200    │
│ 26      │ 'http://localhost:3080/api/keys?name=openAI'                                                                    │ 3503.067000000039  │ 4023.1110000000335 │ 200    │
│ 27      │ 'http://localhost:3080/api/presets'                                                                             │ 3505.585000000021  │ 4023.369000000006  │ 200    │
│ 28      │ 'http://localhost:3080/api/tags'                                                                                │ 3506.645000000077  │ 4023.65200000006   │ 200    │
│ 29      │ 'http://localhost:3080/api/share/link/16390000-0000-4000-8000-000000000001'                                     │ 3506.8790000000736 │ 4247.908000000054  │ 200    │
│ 30      │ 'http://localhost:3080/api/messages/16390000-0000-4000-8000-000000000001?toolPreviews=1'                        │ 3507.116000000038  │ 4538.844000000041  │ 200    │
│ 31      │ 'http://localhost:3080/api/user/settings/favorites/tools'                                                       │ 3511.51800000004   │ 4539.362000000081  │ 200    │
│ 32      │ 'http://localhost:3080/api/endpoints/token-config'                                                              │ 3511.82500000007   │ 4497.655000000028  │ 200    │
│ 33      │ 'http://localhost:3080/api/user/settings/skills/active'                                                         │ 3512.0410000000265 │ 4538.580000000016  │ 200    │
│ 34      │ 'http://localhost:3080/api/agents/tools/web_search/auth'                                                        │ 3512.2420000000275 │ 7276.530000000028  │ 200    │
│ 35      │ 'http://localhost:3080/api/agents/tools/calls?conversationId=16390000-0000-4000-8000-000000000001'              │ 3512.4160000000265 │ 4756.115000000049  │ 200    │
│ 36      │ 'http://localhost:3080/api/agents/chat/status/16390000-0000-4000-8000-000000000001?generationProtocolVersion=2' │ 4673.74700000009   │ 4930.535000000033  │ 200    │
└─────────┴─────────────────────────────────────────────────────────────────────────────────────────────────────────────────┴────────────────────┴────────────────────┴────────┘

Inspect .lighthouse HTML/JSON and e2e/lighthouse/README.md. Reuse loaded user/config data; overlap independent reads without bypassing authorization.

┌─────────┬────────────────────────────┬──────────────────────┬───────┐
│ (index) │ audit                      │ median               │ limit │
├─────────┼────────────────────────────┼──────────────────────┼───────┤
│ 0       │ 'largest-contentful-paint' │ 4557.391             │ 4500  │
│ 1       │ 'cumulative-layout-shift'  │ 0.017901164685666746 │ 0.1   │
│ 2       │ 'total-blocking-time'      │ 416.596              │ 500   │
└─────────┴────────────────────────────┴──────────────────────┴───────┘

  1) [chromium] › e2e/lighthouse/load.spec.ts:10:5 › serial database latency stays within web-vitals budgets 

    Error: Median largest-contentful-paint must stay within 4500

    expect(received).toBeLessThanOrEqual(expected)

    Expected: <= 4500
    Received:    4557.391

       at audit.ts:127

      125 |   console.table(measured);
      126 |   for (const { audit, median, limit } of measured) {
    > 127 |     expect(median, `Median ${audit} must stay within ${limit}`).toBeLessThanOrEqual(limit);
          |                                                                 ^
      128 |   }
      129 |   return results;
      130 | }
        at auditPage (/home/runner/work/LibreChat/LibreChat/e2e/lighthouse/audit.ts:127:65)
        at /home/runner/work/LibreChat/LibreChat/e2e/lighthouse/load.spec.ts:33:19

    attachment #1: screenshot (image/png) ──────────────────────────────────────────────────────────
    e2e/lighthouse/.test-results/load-serial-database-latency-stays-within-web-vitals-budgets-chromium/test-failed-1.png
    ────────────────────────────────────────────────────────────────────────────────────────────────

    Error Context: e2e/lighthouse/.test-results/load-serial-database-latency-stays-within-web-vitals-budgets-chromium/error-context.md

    attachment #3: trace (application/zip) ─────────────────────────────────────────────────────────
    e2e/lighthouse/.test-results/load-serial-database-latency-stays-within-web-vitals-budgets-chromium/trace.zip
    Usage:

        npx playwright show-trace e2e/lighthouse/.test-results/load-serial-database-latency-stays-within-web-vitals-budgets-chromium/trace.zip

    ────────────────────────────────────────────────────────────────────────────────────────────────


🤖: global teardown has been started
2026-11-05 16:37:13 �[32minfo�[39m: �[32mMongo Connection options�[39m
2026-11-05 16:37:13 �[32minfo�[39m: �[32m{�[39m
�[32m  "bufferCommands": false�[39m
�[32m}�[39m
🤖:  ✅  Connected to Database
🤖:  ✅  Found user in Database
🤖:  ✅  Deleted 1 convos & 2 messages
🤖:  ✅  Deleted user from Database
🤖: global teardown has been started
2026-11-05 16:37:14 �[32minfo�[39m: �[32mMongo Connection options�[39m
2026-11-05 16:37:14 �[32minfo�[39m: �[32m{�[39m
�[32m  "bufferCommands": false�[39m
�[32m}�[39m
🤖:  ✅  Connected to Database
🤖:  ⚠️  User not found in Database
  1 failed
    [chromium] › e2e/lighthouse/load.spec.ts:10:5 › serial database latency stays within web-vitals budgets 

Open the full run

[Copilot]

🔵 Needs a closer look

The revocation guarantee depends on MongoDB retries and shared write fences across ACL updates and assignments, requiring confirmation through replica-set concurrency tests that could not run in this review environment.

No open issues 🎉

What changed in this PR

Adds administrator-assigned managers for organization-installed Slack bots, separating bot assignment authority from agent-sharing permission while preserving existing assignments after access is revoked.

Changes:

• Adds tenant-scoped bot-manager grants, revision-checked updates, and transactional assignment authorization.
• Adds an administrator management dialog and limits members’ bot choices to installations they can manage.
• Extends storage, HTTP, component, and browser tests for grants, revocation, replacement, and recovery.

File	Description
packages/​data-schemas/​src/​methods/​slackWorkspace.ts	Integrates grants, filters workspace lists, and clears grants on disconnect.
packages/​data-schemas/​src/​methods/​slackRoute.ts	Checks bot-management authority inside assignment transactions.
packages/​data-schemas/​src/​methods/​slackRoute.spec.ts	Tests grants, revocation races, replacement, pagination, and installation lifecycle.
packages/​data-schemas/​src/​methods/​slackBotAccess.ts	Implements tenant-scoped manager search, reads, and revision-checked writes.
packages/​data-schemas/​src/​admin/​capabilities.ts	Maps Slack bot resources to agent-management capability.
packages/​data-provider/​src/​types/​channels.ts	Adds manager types and bot-management metadata.
packages/​data-provider/​src/​data-service.ts	Adds manager read, search, and update API clients.
packages/​data-provider/​src/​config.ts	Adds the configurable manager limit, defaulting to 50.
packages/​data-provider/​src/​accessPermissions.ts	Defines the Slack bot resource type.
packages/​api/​src/​middleware/​share.ts	Excludes Slack bots from generic sharing.
packages/​api/​src/​middleware/​share.spec.ts	Tests generic-sharing rejection for Slack bots.
packages/​api/​src/​channels/​workspaces.spec.ts	Tests administrator-only manager endpoints and tenant isolation.
packages/​api/​src/​channels/​workspaceRuntime.ts	Supplies the configured manager limit to handlers.
packages/​api/​src/​channels/​workspaceHttp.ts	Adds manager endpoints and permission-aware workspace responses.
packages/​api/​src/​channels/​workspaceBindings.ts	Requires bot-management authority for connection creation and assignment.
packages/​api/​src/​channels/​slackDispatch.spec.ts	Tests manager assignment, forged authority rejection, and post-revocation cleanup.
e2e/​specs/​mock/​scenarios/​slack-workspaces.spec.ts	Exercises manager selection, search recovery, saving, and reopening across themes.
client/​src/​utils/​channelErrors.ts	Maps bot-access failures to localized guidance.
client/​src/​locales/​en/​translation.json	Adds manager-dialog copy and updates assignment guidance.
client/​src/​components/​SidePanel/​Agents/​WorkspaceChannel.tsx	Gates assignment controls on bot access and explains missing permission.
client/​src/​components/​SidePanel/​Agents/​Channels.tsx	Offers only active, manageable bots for new connections.
client/​src/​components/​SidePanel/​Agents/​__tests__/​Channels.spec.tsx	Tests hidden ungranted bots and revoked-access handling.
client/​src/​components/​Channels/​Workspaces.tsx	Adds the administrator’s Manage bot access entry point.
client/​src/​components/​Channels/​BotManagers.tsx	Implements manager selection, removal, saving, and error recovery.
client/​src/​components/​Channels/​BotManagers.spec.tsx	Tests dialog actions, stale saves, and read/search failures.
api/​server/​routes/​slack-workspaces.js	Registers authenticated manager routes with same-origin protection for writes.
api/​server/​controllers/​__tests__/​deleteUserResourceCoverage.spec.js	Records existing principal cleanup as handling bot grants.

_{Review effort: Balanced}

[dustinhealy]

@codex review

[dustinhealy]

Replica-set validation completed locally: all 45 storage cases passed using MongoMemoryReplSet. The revocation test pauses the original database write, commits manager revocation, then releases it; MongoDB retries the conflicted assignment transaction and the revoked grant is denied, leaving no route. Grant refresh, disconnect/reinstallation and tenant-isolation cases also pass. The final HTTP suite additionally verifies that an administrator retains capability-based authority when a redundant direct manager grant is removed during preflight.

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. 🎉

Reviewed commit: 2d87f241e2

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".