Bump agents from 0.14.4 to 0.16.2 in /worker

[dependabot]

Bumps agents from 0.14.4 to 0.16.2.

Release notes

Sourced from agents's releases.

agents@0.16.2

Patch Changes

• #1767 f03dee6 Thanks @​threepointone! - Reduce Think Durable Object SQLite reads during normal wakes and text-only turns.

  Think now avoids automatic media-eviction scans until hydration has been windowed or an oversized appended message has been observed. The shared resumable stream buffer also avoids per-wake metadata-column introspection by creating new tables with the current columns and lazily migrating legacy tables only when a stream write needs it.

• #1722 9f8e14b Thanks @​mattzcarey! - Fix two MCP client OAuth bugs found by the new conformance suite, and add MCP conformance testing.

  • MCPClientConnection now finishes OAuth on the transport that received the 401. A fresh transport loses the resource metadata URL from the WWW-Authenticate header, so token exchange fell back to the default /token path and failed against authorization servers at non-default locations.
  • MCPClientConnection.init() detaches the previous transport before reconnecting. Re-authorizing after a mid-session 401 (scope step-up, token revocation) previously failed permanently with "Already connected to a transport".
  • Added the official @modelcontextprotocol/conformance suite (as used by the MCP TypeScript SDK) running against the MCP client (Agent + MCPClientManager), McpAgent, and createMcpHandler + WorkerTransport — all hosted in workerd via wrangler dev. See packages/agents/conformance/README.md.

• #1767 f03dee6 Thanks @​threepointone! - Cache the active branch tip in AgentSessionProvider so finding the latest leaf no longer scans the whole session on every read and append.

  latestLeafRow() previously ran an anti-join over every message row (O(rows)) to locate the branch tip — on each hydration AND each auto-parent append, so on long transcripts it dominated a wake's read cost. The tip is now maintained in place on append/delete/clear; a cached tip is re-validated on read with an O(1) existence + still-childless check (so it self-heals if another writer deletes the tip or gives it a child), and the full scan only runs when that check fails or the cache is cold. Per-hydration and per-append tip lookups drop from O(rows) to O(1), and the full scan never runs more often than before.

agents@0.16.1

Patch Changes

• #1757 92a5ba1 Thanks @​threepointone! - Bump the partyserver dependency to ^0.5.8, which base64-encodes the x-partykit-props header so props containing non-ASCII characters (e.g. accented names) no longer trigger workerd's "header value contains non-ASCII characters" warning (which throws a TypeError in browser fetch implementations). The header is decoded back to the original Unicode payload on the server, and raw-JSON values from older callers are still accepted for backwards compatibility.

• #1754 151d457 Thanks @​threepointone! - Pin accepted WebSocket connections to binaryType = "arraybuffer". On Worker compatibility_dates >= 2026-03-17 the runtime defaults a server WebSocket's binaryType to "blob" (the websocket_standard_binary_type flag), so binary frames arrive as Blob instead of ArrayBuffer. The Agent protocol and every downstream consumer (e.g. @cloudflare/voice audio frames, user onMessage handlers that check message instanceof ArrayBuffer) have always relied on ArrayBuffer. The Agent now sets connection.binaryType = "arraybuffer" when a connection is established, restoring the historical contract regardless of compatibility date without requiring the no_websocket_standard_binary_type flag. (The hibernatable webSocketMessage handler always delivers ArrayBuffer, so this only affects non-hibernating agents.)

  Also bumps the partyserver dependency to ^0.5.7, which pins binaryType at the connection layer (accept()), accepts non-hibernating connections in half-open mode, and suppresses retryable transport-teardown errors on already-closing/closed connections. With partyserver now pinning binaryType itself, the Agent's own pin becomes defense-in-depth (kept for older partyserver versions and custom connections) and runs once per connection per isolate lifetime instead of on every state access.

• #1754 151d457 Thanks @​threepointone! - Add Browser Run Live View support to the browser tools. The cdp connector gains a getLiveViewUrl({ targetId?, mode? }) tool that returns a link a human

... (truncated)

Changelog

Sourced from agents's changelog.

0.16.2

Patch Changes

• #1767 f03dee6 Thanks @​threepointone! - Reduce Think Durable Object SQLite reads during normal wakes and text-only turns.

  Think now avoids automatic media-eviction scans until hydration has been windowed or an oversized appended message has been observed. The shared resumable stream buffer also avoids per-wake metadata-column introspection by creating new tables with the current columns and lazily migrating legacy tables only when a stream write needs it.

• #1722 9f8e14b Thanks @​mattzcarey! - Fix two MCP client OAuth bugs found by the new conformance suite, and add MCP conformance testing.

  • MCPClientConnection now finishes OAuth on the transport that received the 401. A fresh transport loses the resource metadata URL from the WWW-Authenticate header, so token exchange fell back to the default /token path and failed against authorization servers at non-default locations.
  • MCPClientConnection.init() detaches the previous transport before reconnecting. Re-authorizing after a mid-session 401 (scope step-up, token revocation) previously failed permanently with "Already connected to a transport".
  • Added the official @modelcontextprotocol/conformance suite (as used by the MCP TypeScript SDK) running against the MCP client (Agent + MCPClientManager), McpAgent, and createMcpHandler + WorkerTransport — all hosted in workerd via wrangler dev. See packages/agents/conformance/README.md.

• #1767 f03dee6 Thanks @​threepointone! - Cache the active branch tip in AgentSessionProvider so finding the latest leaf no longer scans the whole session on every read and append.

  latestLeafRow() previously ran an anti-join over every message row (O(rows)) to locate the branch tip — on each hydration AND each auto-parent append, so on long transcripts it dominated a wake's read cost. The tip is now maintained in place on append/delete/clear; a cached tip is re-validated on read with an O(1) existence + still-childless check (so it self-heals if another writer deletes the tip or gives it a child), and the full scan only runs when that check fails or the cache is cold. Per-hydration and per-append tip lookups drop from O(rows) to O(1), and the full scan never runs more often than before.

0.16.1

Patch Changes

• #1757 92a5ba1 Thanks @​threepointone! - Bump the partyserver dependency to ^0.5.8, which base64-encodes the x-partykit-props header so props containing non-ASCII characters (e.g. accented names) no longer trigger workerd's "header value contains non-ASCII characters" warning (which throws a TypeError in browser fetch implementations). The header is decoded back to the original Unicode payload on the server, and raw-JSON values from older callers are still accepted for backwards compatibility.

• #1754 151d457 Thanks @​threepointone! - Pin accepted WebSocket connections to binaryType = "arraybuffer". On Worker compatibility_dates >= 2026-03-17 the runtime defaults a server WebSocket's binaryType to "blob" (the websocket_standard_binary_type flag), so binary frames arrive as Blob instead of ArrayBuffer. The Agent protocol and every downstream consumer (e.g. @cloudflare/voice audio frames, user onMessage handlers that check message instanceof ArrayBuffer) have always relied on ArrayBuffer. The Agent now sets connection.binaryType = "arraybuffer" when a connection is established, restoring the historical contract regardless of compatibility date without requiring the no_websocket_standard_binary_type flag. (The hibernatable webSocketMessage handler always delivers ArrayBuffer, so this only affects non-hibernating agents.)

  Also bumps the partyserver dependency to ^0.5.7, which pins binaryType at the connection layer (accept()), accepts non-hibernating connections in half-open mode, and suppresses retryable transport-teardown errors on already-closing/closed connections. With partyserver now pinning binaryType itself, the Agent's own pin becomes defense-in-depth (kept for older partyserver versions and custom connections) and runs once per connection per isolate lifetime instead of on every state access.

... (truncated)

Commits

• 8019e61 Version Packages (#1768)
• 9f8e14b feat(mcp): MCP conformance test suites + two OAuth client fixes (#1722)
• f03dee6 fix: reduce Think SQLite read amplification (#1767)
• 6f1f71a Version Packages (#1756)
• 92a5ba1 Update Kimi model references to kimi-k2.7-code and bump partyserver to 0.5.8 ...
• 151d457 Browser Run: Live View, Session Recording & Quick Actions + WebSocket/compat ...
• ef85f0a Version Packages (#1708)
• 99c9326 fix(agents): defer agent teardown to its own alarm invocation so canceled req...
• c18a446 fix: stop oversized sessions from bricking the DO with SQLITE_NOMEM on wake (...
• 4ec3b07 fix(agents): ignore RPC responses after socket close (#1748)
• Additional commits viewable in compare view

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.

---

Dependabot commands and options

You can trigger Dependabot actions by commenting on this PR:

• @dependabot rebase will rebase this PR
• @dependabot recreate will recreate this PR, overwriting any edits that have been made to it
• @dependabot show <dependency name> ignore conditions will show all of the ignore conditions of the specified dependency
• @dependabot ignore this major version will close this PR and stop Dependabot creating any more for this major version (unless you reopen the PR or upgrade to it yourself)
• @dependabot ignore this minor version will close this PR and stop Dependabot creating any more for this minor version (unless you reopen the PR or upgrade to it yourself)
• @dependabot ignore this dependency will close this PR and stop Dependabot creating any more for this dependency (unless you reopen the PR or upgrade to it yourself)

[dependabot]

Superseded by #22.