feat(managed-agent): add session lifecycle operations by waleedlatif1 · Pull Request #6140 · simstudioai/sim

waleedlatif1 · 2026-07-31T23:24:23Z

Summary

Adds an operation selector to the Claude Managed Agents block, backed by 9 new tools alongside the existing run-session behavior
Create/send/get/list-events/update/interrupt/archive/delete, plus respond-to-tool-confirmation
A session is now a durable handle addressed by id, so a conversation can span multiple workflow runs instead of being trapped inside one blocking block
get_session resolves blocking tool calls to names/inputs, so an always_ask agent can be approved from a workflow instead of hanging until timeout
Create Session seeds initial_events, collapsing create+send into one API call

Backwards compatibility

Run Session is the default operation. Blocks saved before the selector existed have no stored operation value, so:

the tool selector falls back to managed_agent_run_session
sub-block conditions normalize the missing value to the default, so every field stays visible and required exactly as before
the block's fallback output shape is unchanged (content, sessionId, inputTokens, outputTokens)

blocks/blocks/managed_agent.test.ts pins all of this.

Wire accuracy

Every shape verified against the live Managed Agents docs, not from memory:

user.tool_confirmation uses deny_message (not message) and takes the tool-use event id from stop_reason.event_ids
types[] on the events list is a real repeatable query param
initial_events accepts only user.message/user.define_outcome, validation is all-or-nothing
archive/delete both reject a running session
memory-store endpoints keep the separate agent-memory-2026-07-22 beta header

Type of Change

New feature

Testing

bun run lint, tsc --noEmit, and 183 unit tests pass. Not yet exercised against a live Claude workspace.

Checklist

Code follows project style guidelines
Self-reviewed my changes
Tests added/updated and passing
No new warnings introduced
I confirm that I have read and agree to the terms outlined in the Contributor License Agreement (CLA)

vercel · 2026-07-31T23:24:29Z

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
docs	Ready	Preview	Aug 1, 2026 12:56am

cursor · 2026-07-31T23:24:33Z

PR Summary

Medium Risk
Large new surface area against Anthropic Managed Agents APIs (session lifecycle, irreversible delete/archive, tool-gate wiring); mitigated by extensive unit tests and preserved backward compatibility for existing blocks.

Overview
Extends Claude Managed Agents beyond the single blocking run session path with an operation selector on the block and ten new session tools (create, send message, get session, list events, update, interrupt, archive, delete, respond to tool confirmation, respond to custom tool). Workflows can keep a session id across runs, poll status, read event tails, and unblock requires_action gates instead of hanging on always_ask policies.

The session-client layer gains lifecycle HTTP helpers (retrieveSession, updateSession, archive/delete), initial_events seeding on create, tool confirmations and custom tool results, resolvePendingToolGates with filtered pagination, and newest-first capped event reads with explicit truncated semantics. Docs and integrations.json list all 11 operations; managed_agent.test.ts pins legacy blocks (no stored operation) to managed_agent_run_session and the original field visibility and outputs.

^{Reviewed by Cursor Bugbot for commit 30788ec. Configure here.}

greptile-apps · 2026-07-31T23:27:19Z

Greptile Summary

The PR expands Claude Managed Agents from a blocking run-session tool into a durable session lifecycle API.

Adds create, send, inspect, event-listing, update, interrupt, confirmation, custom-tool response, archive, and delete operations.
Adds operation-specific block inputs, outputs, routing, documentation, and integration metadata while preserving legacy run-session behavior.
Adds session-client pagination, event capping, pending-tool resolution, lifecycle requests, validation, and regression coverage.

Confidence Score: 5/5

The PR appears safe to merge.

No blocking failure remains; the prior documentation, exact-limit truncation, fractional-limit, and non-page-aligned pagination issues are fixed in the current code.

Important Files Changed

Filename	Overview
apps/sim/lib/managed-agents/session-client.ts	Adds durable session lifecycle requests, pending-tool enrichment, and full-history pagination followed by accurate newest-event capping.
apps/sim/tools/managed_agent/list_events.ts	Normalizes event limits, returns capped chronological events, and derives truncation from the untrimmed filtered total.
apps/sim/blocks/blocks/managed_agent.ts	Adds operation selection and operation-specific fields while retaining run-session defaults for existing saved blocks.
apps/docs/content/docs/en/integrations/managed_agent.mdx	Documents the expanded managed-agent lifecycle operations and uses the literal implemented event-limit default.
apps/sim/lib/managed-agents/session-client.test.ts	Covers lifecycle request shapes, pending-tool resolution, pagination boundaries, newest-event retention, totals, and invalid caps.
apps/sim/blocks/blocks/managed_agent.test.ts	Verifies legacy compatibility, operation routing, tool access, field visibility, and required-input behavior.

Sequence Diagram

sequenceDiagram
  participant W as Workflow
  participant B as Managed Agent Block
  participant C as Session Client
  participant A as Anthropic Sessions API
  W->>B: Create session with optional initial message
  B->>C: createSession(...)
  C->>A: POST /v1/sessions
  A-->>W: sessionId
  W->>B: Send message / inspect session
  B->>C: lifecycle operation(sessionId)
  C->>A: Session API request
  A-->>C: status, events, or pending tool gates
  C-->>W: normalized operation output
  alt Tool action required
    W->>B: Confirm tool or return custom result
    B->>C: POST response event
    C->>A: "POST /v1/sessions/{id}/events"
  end
  opt End lifecycle
    W->>B: Interrupt, archive, or delete
    B->>C: lifecycle request
    C->>A: Session state transition
  end

_{Reviews (7): Last reviewed commit: "test(managed-agent): pin the HTTP shape ..." | Re-trigger Greptile}

waleedlatif1 · 2026-07-31T23:39:15Z

@greptile

waleedlatif1 · 2026-07-31T23:39:16Z

@cursor review

waleedlatif1 · 2026-07-31T23:54:55Z

@greptile

waleedlatif1 · 2026-07-31T23:54:56Z

@cursor review

waleedlatif1 · 2026-08-01T00:05:56Z

@greptile

waleedlatif1 · 2026-08-01T00:05:57Z

@cursor review

waleedlatif1 · 2026-08-01T00:18:47Z

@greptile

waleedlatif1 · 2026-08-01T00:18:48Z

@cursor review

waleedlatif1 · 2026-08-01T00:29:44Z

@greptile

waleedlatif1 · 2026-08-01T00:29:45Z

@cursor review

cursor

✅ Bugbot reviewed your changes and found no new issues!

Comment @cursor review or bugbot run to trigger another review on this PR

^{Reviewed by Cursor Bugbot for commit 8ea5162. Configure here.}

Adds an operation selector to the Claude Managed Agents block, backed by nine new tools alongside the existing run-session behavior: - create session (non-blocking, seeds initial_events) - send message to an existing session - get session (surfaces tool calls awaiting approval) - list events - update session (title/metadata) - interrupt session - respond to tool confirmation (allow/deny) - archive session - delete session Run Session stays the default, so blocks saved before the selector existed keep their exact behavior and field layout.

- List Events kept the OLDEST slice when capped, dropping the agent's most recent reply. Paging is exhaustive again and the cap now keeps the newest N after ordering, with a `truncated` flag so callers know it is a tail. - listPaginated returned whole pages past maxItems; it now trims to the exact cap. - A whitespace-only title passed the update guard and would have cleared an existing session title. Blank is now treated as not provided. - Interrupt had no request timeout and could hang; bounded at 15s while still honoring the workflow signal. - Custom-tool gates were surfaced by Get Session but could not be answered, since they need user.custom_tool_result rather than a confirmation. Adds a Respond To Custom Tool operation and a `kind` on each pending gate so a workflow routes to the right one. - Docs rendered a raw ${DEFAULT_EVENT_LIMIT} placeholder for the default.

… result - `truncated` was true whenever the history size equalled the limit, even though nothing was dropped. Event reads now report the untrimmed total and the flag compares against that. - Pending-gate enrichment capped its read, which keeps the OLDEST events in page order — the opposite of where blocking gates live. It now filters to the ids being looked up as pages arrive, which is both correct regardless of page order and bounded by the id count. Paging continues on the raw page so a fully-filtered page is not mistaken for the end of the list. - Respond To Custom Tool applied one result to every id, so multiple pending tools would all receive the same output. It now answers a single call per invocation.

A limit below 1 passed the positivity check and then floored to 0, which made `slice(-0)` hand back the ENTIRE history flagged as complete — the opposite of the requested bound. The limit is now floored before it is validated, so anything that does not resolve to a positive integer falls back to the default. Also hardened the library: a zero or negative cap short-circuits to an empty result instead of falling through to `slice(-0)`, so no future caller can hit the same trap.

The id filter keeps the collected array tiny, so `maxItems` never trips and the walk continued to the end of a session's tool history even after every blocking id had been found. `listPaginated` now takes a `stopWhen` predicate and the gate lookup ends as soon as it has all the ids it came for. Also makes a blocked-but-unnamed session observable: when a session reports `requires_action` with no blocking event ids, `requiresAction` stays true — reporting false would tell a workflow the session is fine while it is parked indefinitely — and the dead end is logged and documented instead.

- A `maxItems` between 0 and 1 slipped past the zero guard and became `slice(-0)` — the whole history — because slice truncates its index toward zero. The cap is now floored at the library boundary, so no caller can hit it whatever they pass. - Update Session documented full metadata replacement but could not express a clear: an empty map normalizes to "absent". Inferring the clear from emptiness would be worse, since an untouched table is also empty and would wipe metadata on every title-only update. Adds an explicit `clearMetadata` instead, and corrects the parameter's documentation.

Method, URL, and beta header for all 11 calls, plus the SSE accept header, the separate memory-store beta (combining the two is a documented 400), and content-type only on requests that carry a body. These are the details types cannot catch and that break silently when a path is "tidied".

waleedlatif1 · 2026-08-01T00:52:28Z

@greptile

waleedlatif1 · 2026-08-01T00:52:29Z

@cursor review

cursor

✅ Bugbot reviewed your changes and found no new issues!

Comment @cursor review or bugbot run to trigger another review on this PR

^{Reviewed by Cursor Bugbot for commit 30788ec. Configure here.}

greptile-apps Bot reviewed Jul 31, 2026

View reviewed changes

Comment thread apps/docs/content/docs/en/integrations/managed_agent.mdx Outdated

cursor Bot reviewed Jul 31, 2026

View reviewed changes

vercel Bot deployed to Preview July 31, 2026 23:28 View deployment

waleedlatif1 force-pushed the feat/managed-agents-endpoint-audit branch from 80d79f3 to 0adfa65 Compare July 31, 2026 23:38

greptile-apps Bot reviewed Jul 31, 2026

View reviewed changes

Comment thread apps/sim/tools/managed_agent/list_events.ts Outdated

cursor Bot reviewed Jul 31, 2026

View reviewed changes

Comment thread apps/sim/lib/managed-agents/session-client.ts

Comment thread apps/sim/tools/managed_agent/list_events.ts Outdated

Comment thread apps/sim/tools/managed_agent/respond_custom_tool.ts Outdated

vercel Bot deployed to Preview July 31, 2026 23:44 View deployment

waleedlatif1 force-pushed the feat/managed-agents-endpoint-audit branch from 0adfa65 to 4e991d0 Compare July 31, 2026 23:54

greptile-apps Bot reviewed Jul 31, 2026

View reviewed changes

Comment thread apps/sim/tools/managed_agent/list_events.ts Outdated

vercel Bot deployed to Preview July 31, 2026 23:58 View deployment

cursor Bot reviewed Jul 31, 2026

View reviewed changes

Comment thread apps/sim/lib/managed-agents/session-client.ts

vercel Bot temporarily deployed to Preview August 1, 2026 00:05 Inactive

cursor Bot reviewed Aug 1, 2026

View reviewed changes

Comment thread apps/sim/lib/managed-agents/session-client.ts

Comment thread apps/sim/tools/managed_agent/get_session.ts

waleedlatif1 force-pushed the feat/managed-agents-endpoint-audit branch from b4807ca to 1f1ec2d Compare August 1, 2026 00:18

cursor Bot reviewed Aug 1, 2026

View reviewed changes

Comment thread apps/sim/lib/managed-agents/session-client.ts Outdated

Comment thread apps/sim/tools/managed_agent/update_session.ts

vercel Bot deployed to Preview August 1, 2026 00:23 View deployment

cursor Bot reviewed Aug 1, 2026

View reviewed changes

vercel Bot deployed to Preview August 1, 2026 00:33 View deployment

vercel Bot temporarily deployed to Preview August 1, 2026 00:51 Inactive

waleedlatif1 added 7 commits July 31, 2026 17:51

waleedlatif1 force-pushed the feat/managed-agents-endpoint-audit branch from 398229e to 30788ec Compare August 1, 2026 00:52

cursor Bot reviewed Aug 1, 2026

View reviewed changes

vercel Bot deployed to Preview August 1, 2026 00:56 View deployment

waleedlatif1 merged commit 19b0312 into staging Aug 1, 2026
27 checks passed

waleedlatif1 deleted the feat/managed-agents-endpoint-audit branch August 1, 2026 00:59

waleedlatif1 mentioned this pull request Aug 1, 2026

v0.7.51: hybrid search for knowledgebases, ui improvements, db perf improvements, claude managed agents tools, realtime files and tables #6146

Merged

Uh oh!

Conversation

waleedlatif1 commented Jul 31, 2026

Summary

Backwards compatibility

Wire accuracy

Type of Change

Testing

Checklist

Uh oh!

vercel Bot commented Jul 31, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

cursor Bot commented Jul 31, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

PR Summary

Uh oh!

greptile-apps Bot commented Jul 31, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Greptile Summary

Confidence Score: 5/5

Important Files Changed

Sequence Diagram

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

waleedlatif1 commented Jul 31, 2026

Uh oh!

waleedlatif1 commented Jul 31, 2026

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

waleedlatif1 commented Jul 31, 2026

Uh oh!

waleedlatif1 commented Jul 31, 2026

Uh oh!

Uh oh!

Uh oh!

waleedlatif1 commented Aug 1, 2026

Uh oh!

waleedlatif1 commented Aug 1, 2026

Uh oh!

Uh oh!

Uh oh!

waleedlatif1 commented Aug 1, 2026

Uh oh!

waleedlatif1 commented Aug 1, 2026

Uh oh!

Uh oh!

Uh oh!

waleedlatif1 commented Aug 1, 2026

Uh oh!

waleedlatif1 commented Aug 1, 2026

Uh oh!

cursor Bot left a comment

Choose a reason for hiding this comment

Uh oh!

waleedlatif1 commented Aug 1, 2026

Uh oh!

waleedlatif1 commented Aug 1, 2026

Uh oh!

cursor Bot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

vercel Bot commented Jul 31, 2026 •

edited

Loading

cursor Bot commented Jul 31, 2026 •

edited

Loading

greptile-apps Bot commented Jul 31, 2026 •

edited

Loading