Merge feat/sub-agents: wave 4 — job abort (real kill + abort steer), job usage capture, JobUsage.lastStep

This commit is contained in:
amosblomqvist
2026-09-11 03:27:58 +03:00
10 changed files with 644 additions and 33 deletions
+13 -1
View File
@@ -49,7 +49,7 @@ import { runDemoBuilder } from "../src/agent/subagents/demo-builder.ts";
import { headlessOverride, type DemoRunResult, type HeadlessApi } from "../src/agent/subagents/headless.ts";
import { createJobRunner, demoModelOverride, visualModelOverride } from "../src/agent/subagents/runner.ts";
import { runVisualMaker } from "../src/agent/subagents/visual-maker.ts";
import { buildSubagentSteer, isSubagentFailure, type SubagentSteerMessage } from "../src/core/contracts.ts";
import { buildAbortSteer, buildSubagentSteer, isSubagentFailure, type SubagentSteerMessage } from "../src/core/contracts.ts";
import type { JobRecord } from "../src/core/contracts.ts";
import { createTeacherTools } from "../src/agent/tools.ts";
import { chatStore } from "../src/store/chats.ts";
@@ -342,6 +342,18 @@ describe("steer card rendering (the delivered artifact)", () => {
assert.ok(!html.includes("Looking at the render"), "model-facing notes never render");
});
it("renders the abort chip for an aborted job's steer (2026-09 dock semantics)", () => {
const steer = buildAbortSteer({ id: "jc", kind: "research", toolCallId: "cc" });
assert.equal(steer.metadata.subagentResult.aborted, true);
assert.equal(steer.parts.length, 1, "no artifact part on an abort");
const html = renderToStaticMarkup(<SubagentSteerCard message={steer as never} />);
assert.match(html, /job-aborted-badge/);
assert.match(html, /research/, "the job's kind, lower-cased like the complete-badge idiom");
assert.match(html, /aborted by the learner/);
// the teacher-facing protocol text never renders for the learner
assert.ok(!html.includes("no brief is coming"), "no protocol glue in the chip");
});
it("renders a quiet Research-complete one-liner whose collapsed <details> carries the full brief", () => {
const steer = buildSubagentSteer(
{ id: "jr", kind: "research", toolCallId: "cr" },
+246 -1
View File
@@ -31,18 +31,23 @@ import type { UIMessage } from "ai";
import { MockLanguageModelV4, simulateReadableStream } from "ai/test";
import type { LanguageModelV4StreamPart } from "@ai-sdk/provider";
import {
accumulateJobUsage,
buildAbortSteer,
buildSubagentSteer,
isSubagentFailure,
steerDelivered,
type DemoResult,
type JobUsage,
type ResearchResult,
type SteerMessage,
type SubagentSteerMessage,
type VisualResult,
} from "../src/core/contracts.ts";
import { composeInstructions, createTeacherAgent } from "../src/agent/teacher-agent.ts";
import { createTeacherTools } from "../src/agent/tools.ts";
import { handleChatRequest } from "../src/server/chat-handler.ts";
import { claimSteers, deliverJobResult, setTurnRunner } from "../src/server/jobs.ts";
import { abortJob, claimSteers, deliverJobResult, setTurnRunner } from "../src/server/jobs.ts";
import { beginTurn, endTurn } from "../src/server/turn-liveness.ts";
import { persistTurn, withMessageIds } from "../src/server/turn-service.ts";
import { chatStore } from "../src/store/chats.ts";
import { jobStore } from "../src/store/jobs.ts";
@@ -170,6 +175,57 @@ describe("steer helpers (contracts.ts)", () => {
assert.deepEqual(data.data.result, research);
});
it("accumulateJobUsage: sums per model call; fields NO call reported stay undefined (absent is not zero)", () => {
let total: JobUsage | undefined = undefined;
total = accumulateJobUsage(total, "z-ai/glm-5.3-flash", {
inputTokens: 10,
outputTokens: 5,
inputTokenDetails: { cacheReadTokens: 4 },
});
assert.deepEqual(total, {
model: "z-ai/glm-5.3-flash",
inputTokens: 10,
outputTokens: 5,
cacheReadTokens: 4,
cacheWriteTokens: undefined,
reasoningTokens: undefined,
steps: 1,
lastStep: {
inputTokens: 10,
outputTokens: 5,
cacheReadTokens: 4,
cacheWriteTokens: undefined,
reasoningTokens: undefined,
},
}, "lastStep is the most recent call's accounting, not cumulative");
// a later call that reports nothing new keeps reported fields summing
// (unreported contributes 0) and bumps the billing step count.
total = accumulateJobUsage(total, "z-ai/glm-5.3-flash", { inputTokens: 20, outputTokens: 0 });
assert.equal(total.inputTokens, 30);
assert.equal(total.outputTokens, 5);
assert.equal(total.cacheReadTokens, 4, "unreported after reported contributes 0");
assert.equal(total.cacheWriteTokens, undefined, "never reported stays absent — never zero");
assert.equal(total.reasoningTokens, undefined);
assert.equal(total.steps, 2, "one step per model call, reported or not");
assert.deepEqual(total.lastStep, {
inputTokens: 20,
outputTokens: 0,
cacheReadTokens: undefined,
cacheWriteTokens: undefined,
reasoningTokens: undefined,
}, "lastStep always mirrors the LATEST call, never the sums");
});
it("buildAbortSteer: metadata-aborted, text-only (no artifact), same deterministic id space as the result steer", () => {
const steer = buildAbortSteer({ id: "job_c", kind: "demo", toolCallId: "call-c" });
assert.equal(steer.id, "steer_job_c", "same id space — steerDelivered treats either as delivered");
assert.equal(steer.role, "user");
assert.equal(steer.metadata.subagentResult.aborted, true);
assert.equal(steer.metadata.subagentResult.jobId, "job_c");
assert.equal(steer.parts.length, 1);
assert.ok(steer.parts[0]!.type === "text" && /aborted/.test(steer.parts[0].text));
});
it("isSubagentFailure: failure shapes only (empty html/svg, failure brief)", () => {
assert.equal(isSubagentFailure("research", research), false);
assert.equal(isSubagentFailure("research", { brief: "Research failed: x", sources: [] }), true);
@@ -836,3 +892,192 @@ describe("prompt composition", () => {
assert.ok(composeInstructions().length <= 60_000, "composed system prompt over the 60k guard");
});
});
// ─── abort (2026-09 semantics: real kill, cancelled status, teacher told) ───
describe("abort (src/server/jobs.ts abortJob)", () => {
/** A hanging model that only settles when its abort signal fires — the
* real-kill path: the in-flight call itself is killed. */
function hangingModel(): MockLanguageModelV4 {
return new MockLanguageModelV4({
doGenerate: ({ abortSignal }) =>
new Promise((_resolve, reject) => {
// an ALREADY-aborted signal never fires the event — real
// providers reject an aborted fetch at once, the mock must too
if (abortSignal?.aborted) {
reject(new DOMException("The operation was aborted", "AbortError"));
return;
}
abortSignal?.addEventListener("abort", () => {
reject(new DOMException("The operation was aborted", "AbortError"));
});
}),
});
}
/** One step (with usage) then a hang — proves usage-so-far survives the kill. */
function oneStepThenHangModel(): MockLanguageModelV4 {
let turn = 0;
return new MockLanguageModelV4({
doGenerate: ({ abortSignal }) => {
turn++;
if (turn === 1) {
return Promise.resolve({
content: [
{
type: "tool-call" as const,
toolCallId: "ws1",
toolName: "web_search",
input: JSON.stringify({ query: "anything" }),
},
],
finishReason: { unified: "tool-calls" as const, raw: "tool-calls" },
usage: usage(),
warnings: [],
});
}
return new Promise((_resolve, reject) => {
if (abortSignal?.aborted) {
reject(new DOMException("The operation was aborted", "AbortError"));
return;
}
abortSignal?.addEventListener("abort", () => {
reject(new DOMException("The operation was aborted", "AbortError"));
});
});
},
});
}
function newChatWithResearchJob(label: string, toolCallId: string): { chatId: string; jobId: string } {
const chat = chatStore.createChat(`Test: ${label}`);
trackChat(chat.id);
chatStore.saveMessages(chat.id, [
{ id: "u1", role: "user", parts: [{ type: "text", text: "hi" }] },
{
id: "a1",
role: "assistant",
parts: [
{ type: "text", text: "Checking." },
{ type: "tool-research", toolCallId, state: "output-available", input: RESEARCH_INPUT, output: "Dispatched." },
],
},
] as never);
const job = jobStore.createJob({ kind: "research", chatId: chat.id, toolCallId });
trackJob(job.id);
return { chatId: chat.id, jobId: job.id };
}
it("kills the running loop: error \"Aborted by learner\" immediately, loop unwinds silently, teacher gets the abort steer", async () => {
const { jobId } = newChatWithResearchJob("abort running job", "call-abort1");
const fired: SteerMessage[] = [];
setTurnRunner(async (_cid, messages) => {
fired.push(messages[messages.length - 1] as SteerMessage);
return 200;
});
researchModelOverride.current = hangingModel();
const runner = createJobRunner({
chatStore,
jobStore,
deliver: async () => {
throw new Error("a cancelled job must never deliver a result steer");
},
});
try {
const settled = runner.runJob(jobId, RESEARCH_INPUT);
await waitFor(() => jobStore.getJob(jobId)!.activity === "researching", 8000, "loop started");
const outcome = abortJob(jobId);
assert.ok(outcome.ok);
const killed = jobStore.getJob(jobId)!;
assert.equal(killed.status, "error", "flipped synchronously — the poll sees it at once");
assert.equal(killed.error, "Aborted by learner", "abort reads as an error with a reason, not a failure");
assert.equal(killed.activity, "aborted");
await settled; // unwinds — no hang
assert.equal(jobStore.getJob(jobId)!.status, "error", "the runner never overwrites the flip");
assert.equal(jobStore.getJob(jobId)!.error, "Aborted by learner");
assert.equal(jobStore.getJob(jobId)!.result, undefined, "no result on an aborted job");
await waitFor(() => fired.length > 0, 8000, "abort wake fired");
const last = fired[0]!;
assert.equal(last.metadata.subagentResult.aborted, true);
assert.equal(last.id, `steer_${jobId}`);
assert.equal(last.parts.length, 1, "text only — no artifact part");
// aborted is terminal: the wake completing must not flip it
assert.equal(jobStore.getJob(jobId)!.status, "error", "terminal after the wake");
} finally {
researchModelOverride.current = undefined;
}
});
it("the usage folded so far survives the kill (money channel outlives the abort)", async () => {
const restoreFetch = stubFetch();
try {
const { jobId } = newChatWithResearchJob("abort keeps usage", "call-abort2");
setTurnRunner(async () => 200);
researchModelOverride.current = oneStepThenHangModel();
const runner = createJobRunner({ chatStore, jobStore, deliver: async () => {} });
try {
const settled = runner.runJob(jobId, RESEARCH_INPUT);
await waitFor(() => (jobStore.getJob(jobId)!.usage?.steps ?? 0) === 1, 8000, "first step folded");
abortJob(jobId);
await settled;
const job = jobStore.getJob(jobId)!;
assert.equal(job.status, "error");
assert.equal(job.error, "Aborted by learner");
assert.equal(job.usage!.steps, 1, "usage-so-far persisted by the runner's finally");
assert.equal(job.usage!.model, "z-ai/glm-5.3-flash", "the config role's OpenRouter id");
assert.equal(job.usage!.inputTokens, 10);
assert.equal(job.usage!.outputTokens, 5);
} finally {
researchModelOverride.current = undefined;
}
} finally {
restoreFetch();
}
});
it("aborting a delivering job drops the queued result steer and wakes with the abort note instead", async () => {
const chat = chatStore.createChat("Test: abort delivering job");
trackChat(chat.id);
chatStore.saveMessages(chat.id, [{ id: "u1", role: "user", parts: [{ type: "text", text: "hello" }] }] as never);
const job = jobStore.createJob({ kind: "research", chatId: chat.id, toolCallId: "call-abort3" });
trackJob(job.id);
const result: ResearchResult = { brief: "Done: it works.", sources: [{ title: "s", url: "https://x.test" }] };
jobStore.updateJob(job.id, { result, activity: "delivering result" });
// Block the pump's idle gates so the result steer sits QUEUED, not in flight.
assert.ok(beginTurn(chat.id));
const fired: SteerMessage[] = [];
setTurnRunner(async (_cid, messages) => {
fired.push(messages[messages.length - 1] as SteerMessage);
return 200;
});
void deliverJobResult(buildSubagentSteer(job, result)).catch(() => {
/* dropped by the abort below */
});
await waitFor(() => jobStore.getJob(job.id)!.activity === "delivering result", 8000, "queued");
const outcome = abortJob(job.id);
assert.ok(outcome.ok);
assert.equal(jobStore.getJob(job.id)!.status, "error");
assert.equal(jobStore.getJob(job.id)!.error, "Aborted by learner");
endTurn(chat.id); // gates open — whatever is left in the queue fires
await waitFor(() => fired.length > 0, 8000, "wake fired");
assert.equal(fired.length, 1, "exactly one delivery — the result steer was dropped");
assert.equal(fired[0]!.metadata.subagentResult.aborted, true, "the abort steer, not the result");
assert.equal(jobStore.getJob(job.id)!.status, "error", "terminal after the wake");
});
it("outcome reasons: not_found for unknown, not_running for settled jobs", () => {
assert.deepEqual(abortJob("job_does_not_exist"), { ok: false, reason: "not_found" });
const chat = chatStore.createChat("Test: abort settled job");
trackChat(chat.id);
const job = jobStore.createJob({ kind: "research", chatId: chat.id, toolCallId: "call-abort4" });
trackJob(job.id);
jobStore.updateJob(job.id, { status: "done", activity: "delivered" });
assert.deepEqual(abortJob(job.id), { ok: false, reason: "not_running" });
});
});