fix(ai): preserve Vertex Anthropic tool continuations (#43498)

Signed-off-by: Major Hayden <major@mhtx.net>
This commit is contained in:
Major Hayden 2026-08-25 17:22:51 -05:00 committed by GitHub
parent c2a3b813a0
commit 2e7f06a155
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
2 changed files with 157 additions and 4 deletions

View file

@ -744,9 +744,12 @@ const endsInServerToolUse = (message: LLMRequest["messages"][number]) => {
return message.role === "assistant" && last?.type === "tool-call" && last.providerExecuted === true
}
const canUseNativeSystemUpdate = (messages: LLMRequest["messages"], index: number) => {
const previous = messages[index - 1]
const next = messages[index + 1]
const canUseNativeSystemUpdate = (request: LLMRequest, index: number) => {
const previous = request.messages[index - 1]
const next = request.messages[index + 1]
// Vertex currently rejects/404s for a system message after local tool results,
// so fold it into the user tool-result turn across continuations and history.
if (request.model.route.id === "google-vertex-messages" && previous?.role === "tool") return false
return (
previous !== undefined &&
previous.role !== "system" &&
@ -793,7 +796,7 @@ const lowerMessages = Effect.fn("AnthropicMessages.lowerMessages")(function* (
if (message.role === "system") {
if (splitsLocalToolResults(request.messages, index))
return yield* invalid("Anthropic Messages system updates cannot split a local tool call from its tool result")
if (supportsNativeSystemUpdates(request) && canUseNativeSystemUpdate(request.messages, index)) {
if (supportsNativeSystemUpdates(request) && canUseNativeSystemUpdate(request, index)) {
messages.push(yield* lowerNativeSystemUpdate(message, breakpoints))
continue
}

View file

@ -5,6 +5,7 @@ import { CacheHint, LLM, AIError, LLMRequest, Message, ToolCallPart, ToolDefinit
import { Auth, LLMClient } from "../../src/route.js"
import { compileRequest } from "../../src/route/client.js"
import * as AnthropicMessages from "../../src/protocols/anthropic-messages.js"
import { GoogleVertexMessages } from "../../src/providers.js"
import { continuationRequest, nativeAnthropicMessagesContinuation } from "../continuation-scenarios.js"
import { it } from "../lib/effect.js"
import { dynamicResponse, fixedResponse } from "../lib/http.js"
@ -27,6 +28,12 @@ const compileUnsignedReasoning = (model: LLMRequest["model"]) =>
}),
)
const vertexOpus48 = GoogleVertexMessages.configure({
accessToken: "test",
location: "global",
project: "test",
}).model("claude-opus-4-8")
const request = LLM.request({
id: "req_1",
model,
@ -286,6 +293,149 @@ describe("Anthropic Messages route", () => {
}),
)
it.effect("keeps a terminal Vertex system update in the tool-result turn", () =>
Effect.gen(function* () {
const prepared = yield* compileRequest(
LLM.request({
model: vertexOpus48,
messages: [
Message.assistant([ToolCallPart.make({ id: "call_1", name: "lookup", input: {} })]),
Message.tool({ id: "call_1", name: "lookup", result: "Done." }),
Message.system("Operator update."),
],
cache: "none",
}),
)
expect(prepared.body.messages).toEqual([
{
role: "assistant",
content: [{ type: "tool_use", id: "call_1", name: "lookup", input: {} }],
},
{
role: "user",
content: [
{
type: "tool_result",
tool_use_id: "call_1",
content: '"Done."',
is_error: undefined,
cache_control: undefined,
},
{
type: "text",
text: "<system-update>\nOperator update.\n</system-update>",
cache_control: undefined,
},
],
},
])
}),
)
it.effect("preserves folded tool-result system updates across multi-turn Vertex history", () =>
Effect.gen(function* () {
const prepared = yield* compileRequest(
LLM.request({
model: vertexOpus48,
messages: [
Message.assistant([ToolCallPart.make({ id: "call_1", name: "lookup", input: {} })]),
Message.tool({ id: "call_1", name: "lookup", result: "Done." }),
Message.system("Operator update."),
Message.assistant("Acknowledged."),
Message.user("Next step."),
],
cache: "none",
}),
)
expect(prepared.body.messages).toEqual([
{
role: "assistant",
content: [{ type: "tool_use", id: "call_1", name: "lookup", input: {} }],
},
{
role: "user",
content: [
{
type: "tool_result",
tool_use_id: "call_1",
content: '"Done."',
is_error: undefined,
cache_control: undefined,
},
{
type: "text",
text: "<system-update>\nOperator update.\n</system-update>",
cache_control: undefined,
},
],
},
{ role: "assistant", content: [{ type: "text", text: "Acknowledged." }] },
{ role: "user", content: [{ type: "text", text: "Next step." }] },
])
}),
)
it.effect("keeps a terminal direct Anthropic system update native", () =>
Effect.gen(function* () {
const prepared = yield* compileRequest(
LLM.request({
model: opus48,
messages: [
Message.assistant([ToolCallPart.make({ id: "call_1", name: "lookup", input: {} })]),
Message.tool({ id: "call_1", name: "lookup", result: "Done." }),
Message.system("Operator update."),
],
cache: "none",
}),
)
expect(prepared.body.messages).toEqual([
{
role: "assistant",
content: [{ type: "tool_use", id: "call_1", name: "lookup", input: {} }],
},
{
role: "user",
content: [
{
type: "tool_result",
tool_use_id: "call_1",
content: '"Done."',
is_error: undefined,
cache_control: undefined,
},
],
},
{
role: "system",
content: [{ type: "text", text: "Operator update.", cache_control: undefined }],
},
])
}),
)
it.effect("keeps an ordinary terminal Vertex system update native", () =>
Effect.gen(function* () {
const prepared = yield* compileRequest(
LLM.request({
model: vertexOpus48,
messages: [Message.user("Before."), Message.system("Operator update.")],
cache: "none",
}),
)
expect(prepared.body.messages).toEqual([
{ role: "user", content: [{ type: "text", text: "Before." }] },
{
role: "system",
content: [{ type: "text", text: "Operator update.", cache_control: undefined }],
},
])
}),
)
it.effect("rejects a system update between a local tool call and its result", () =>
Effect.gen(function* () {
const error = yield* compileRequest(