Files
pikasTech-HWLAB/scripts/code-agent-chat-smoke.mjs
T
Lyon 07e4638264 fix: show Code Agent completion evidence
Host commander review: merged PR #171 after diff review. Scope is user-visible Code Agent completion evidence and smoke guards that reject echo/mock/stub completions; live smoke is read-only aside from the minimal chat request and does not print reply text or secrets. No runtime mutation, no service restart, and no M3/M4/M5 acceptance claim.
2026-05-23 02:15:21 +08:00

386 lines
13 KiB
JavaScript

#!/usr/bin/env node
import assert from "node:assert/strict";
import http from "node:http";
import https from "node:https";
import {
handleCodeAgentChat,
validateCodeAgentChatSchema
} from "../internal/cloud/code-agent-chat.mjs";
const defaultLiveUrl = "http://74.48.78.17:16666/";
const defaultLiveMessage = "请用一句话回复 HWLAB Code Agent readiness。";
const trustedLiveProviders = new Set(["openai-responses", "codex-cli"]);
const untrustedProviderPattern = /\b(?:echo|mock|fixture|stub|sample)\b/iu;
const args = parseArgs(process.argv.slice(2));
if (args.mode === "help") {
process.stdout.write(`${JSON.stringify(usage(), null, 2)}\n`);
} else if (args.mode === "live") {
const report = await runLiveReadinessSmoke(args);
process.stdout.write(`${JSON.stringify(report, null, 2)}\n`);
process.exitCode = report.status === "pass" ? 0 : 2;
} else {
await runLocalContractSmoke();
}
function logOk(name) {
process.stdout.write(`[code-agent-chat-smoke] ok ${name}\n`);
}
async function runLocalContractSmoke() {
const completed = await handleCodeAgentChat(
{
conversationId: "cnv_code-agent-chat-smoke",
traceId: "trc_code-agent-chat-smoke",
projectId: "prj_code-agent-chat-smoke",
message: "请用一句话说明当前 HWLAB 工作台可以做什么。"
},
{
now: () => "2026-05-22T00:00:00.000Z",
callProvider: async ({ providerPlan }) => ({
provider: providerPlan.provider,
model: providerPlan.model,
backend: providerPlan.backend,
content: "HWLAB 工作台可以用 Code Agent 对话来整理任务、查看资源并保持硬件变更受控。",
usage: null,
providerTrace: {
source: "schema-smoke-provider"
}
})
}
);
validateCodeAgentChatSchema(completed);
assert.equal(completed.conversationId, "cnv_code-agent-chat-smoke");
assert.equal(completed.sessionId, "cnv_code-agent-chat-smoke");
assert.match(completed.messageId, /^msg_/);
assert.equal(completed.status, "completed");
assert.equal(completed.createdAt, "2026-05-22T00:00:00.000Z");
assert.equal(completed.updatedAt, "2026-05-22T00:00:00.000Z");
assert.equal(completed.traceId, "trc_code-agent-chat-smoke");
assert.equal(completed.error, undefined);
assert.match(completed.reply.content, /Code Agent/);
logOk("completed schema");
const sourceReadiness = classifyCodeAgentChatReadiness(completed, { realDevLive: false });
assert.equal(sourceReadiness.status, "blocked");
assert.equal(sourceReadiness.level, "SOURCE");
assert.equal(sourceReadiness.devLiveReplyPass, false);
logOk("local stub completion is not #143 DEV-LIVE pass");
const echoReadiness = classifyCodeAgentChatReadiness({
...completed,
provider: "echo-mock",
model: "mock-model",
backend: "hwlab-cloud-api/echo-mock"
}, { realDevLive: true, httpStatus: 200 });
assert.equal(echoReadiness.status, "blocked");
assert.equal(echoReadiness.level, "BLOCKED/untrusted-completion");
assert.equal(echoReadiness.blocker, "untrusted-completion");
assert.equal(echoReadiness.devLiveReplyPass, false);
assert.match(echoReadiness.reason, /echo\/mock\/stub/u);
logOk("echo/mock completion is not real DEV-LIVE pass");
const failed = await handleCodeAgentChat(
{
traceId: "trc_code-agent-chat-smoke-failed",
message: "你好"
},
{
now: () => "2026-05-22T00:01:00.000Z",
env: {
PATH: "",
HWLAB_CODE_AGENT_PROVIDER: "codex-cli",
HWLAB_CODE_AGENT_MODEL: "gpt-test"
}
}
);
validateCodeAgentChatSchema(failed);
assert.match(failed.conversationId, /^cnv_/);
assert.match(failed.sessionId, /^cnv_/);
assert.match(failed.messageId, /^msg_/);
assert.equal(failed.status, "failed");
assert.equal(failed.createdAt, "2026-05-22T00:01:00.000Z");
assert.equal(failed.updatedAt, "2026-05-22T00:01:00.000Z");
assert.equal(failed.traceId, "trc_code-agent-chat-smoke-failed");
assert.equal(failed.provider, "codex-cli");
assert.equal(failed.model, "gpt-test");
assert.equal(failed.backend, "hwlab-cloud-api/codex-cli");
assert.equal(failed.error.code, "provider_unavailable");
assert.match(failed.error.message, /Codex CLI command is not available/);
assert.deepEqual(failed.error.missingCommands, ["codex"]);
assert.ok(failed.error.missingEnv.includes("OPENAI_API_KEY"));
assert.equal(failed.availability.status, "blocked");
assert.match(failed.availability.blocker, /凭证缺口|OPENAI_API_KEY/u);
assert.equal(failed.availability.reason, "provider_unavailable");
assert.match(failed.availability.summary, /真实后端已接入/u);
assert.match(failed.availability.summary, /hwlab-code-agent-provider\/openai-api-key/u);
assert.equal(JSON.stringify(failed).includes("sk-"), false);
assert.equal(Object.hasOwn(failed, "reply"), false);
logOk("failed provider-gap schema");
const credentialReadiness = classifyCodeAgentChatReadiness(failed, { realDevLive: true });
assert.equal(credentialReadiness.status, "blocked");
assert.equal(credentialReadiness.level, "BLOCKED/credential");
assert.equal(credentialReadiness.blocker, "credential");
assert.equal(credentialReadiness.devLiveReplyPass, false);
logOk("provider credential blocker readiness");
process.stdout.write("[code-agent-chat-smoke] passed\n");
}
async function runLiveReadinessSmoke(args) {
const endpoint = agentChatEndpoint(args.url);
const traceId = protocolId("trc_code-agent-chat-live");
const conversationId = protocolId("cnv_code-agent-chat-live");
let httpStatus = null;
let payload = null;
let transportError = null;
try {
const response = await postJson(endpoint, {
headers: {
"x-trace-id": traceId
},
body: {
conversationId,
traceId,
projectId: "prj_hwlab-cloud-workbench",
message: args.message
}
});
httpStatus = response.status;
payload = response.json;
} catch (error) {
transportError = error instanceof Error ? error.message : String(error);
}
const readiness = payload
? classifyCodeAgentChatReadiness(payload, { realDevLive: true, httpStatus })
: {
status: "blocked",
level: "BLOCKED",
blocker: "transport",
devLiveReplyPass: false,
reason: "real DEV /v1/agent/chat did not return a JSON Code Agent payload"
};
if (payload) {
try {
validateCodeAgentChatSchema(payload);
} catch (error) {
readiness.status = "blocked";
readiness.level = "BLOCKED";
readiness.blocker = "schema";
readiness.devLiveReplyPass = false;
readiness.reason = error instanceof Error ? error.message : String(error);
}
}
return {
status: readiness.status,
mode: "live-readiness",
url: endpoint.href,
httpStatus,
readiness,
response: summarizePayload(payload),
...(transportError ? { transportError } : {}),
safety: {
prodTouched: false,
servicesRestarted: false,
readsSecretValues: false,
printsAssistantReply: false,
statement: "This smoke posts only to the configured DEV Code Agent chat endpoint and prints reply presence, not reply content or secret values."
}
};
}
function classifyCodeAgentChatReadiness(payload, { realDevLive = false, httpStatus = null } = {}) {
const assistantReply = typeof payload?.reply?.content === "string" ? payload.reply.content.trim() : "";
if (payload?.status === "completed" && assistantReply) {
const evidence = inspectRealCompletionEvidence(payload);
if (!evidence.ok) {
return {
status: "blocked",
level: realDevLive ? "BLOCKED/untrusted-completion" : "SOURCE",
blocker: "untrusted-completion",
devLiveReplyPass: false,
reason: `completed response is missing real provider/model/trace/conversation evidence or looks like echo/mock/stub: ${evidence.reason}`
};
}
if (realDevLive) {
return {
status: "pass",
level: "#143 DEV-LIVE reply pass",
blocker: null,
devLiveReplyPass: true,
reason: "real DEV /v1/agent/chat completed with a non-empty assistant reply"
};
}
return {
status: "blocked",
level: "SOURCE",
blocker: "not-dev-live",
devLiveReplyPass: false,
reason: "mock, fixture, local stub, or source-only completion cannot be counted as #143 DEV-LIVE reply pass"
};
}
const missingEnv = Array.isArray(payload?.error?.missingEnv) ? payload.error.missingEnv : [];
if (
payload?.status === "failed" &&
payload.error?.code === "provider_unavailable" &&
(missingEnv.includes("OPENAI_API_KEY") || missingEnv.includes("HWLAB_CODE_AGENT_OPENAI_BASE_URL"))
) {
return {
status: "blocked",
level: "BLOCKED/credential",
blocker: missingEnv.includes("OPENAI_API_KEY") ? "credential" : "provider-config",
devLiveReplyPass: false,
reason: "provider_unavailable with missing Code Agent provider env means the provider contract is incomplete"
};
}
return {
status: "blocked",
level: "BLOCKED",
blocker: "runtime",
devLiveReplyPass: false,
reason: `Code Agent chat did not produce a completed non-empty assistant reply${httpStatus ? ` (HTTP ${httpStatus})` : ""}`
};
}
function inspectRealCompletionEvidence(payload) {
const provider = String(payload?.provider ?? "").trim().toLowerCase();
const backend = String(payload?.backend ?? "").trim().toLowerCase();
const missing = [];
if (!payload?.conversationId && !payload?.sessionId) missing.push("conversationId");
if (!payload?.traceId) missing.push("traceId");
if (!payload?.messageId) missing.push("messageId");
if (!payload?.provider) missing.push("provider");
if (!payload?.model) missing.push("model");
if (!payload?.backend) missing.push("backend");
if (missing.length > 0) {
return { ok: false, reason: `missing ${missing.join(",")}` };
}
if (!trustedLiveProviders.has(provider)) {
return { ok: false, reason: `provider ${payload.provider} is not an accepted real provider` };
}
if (untrustedProviderPattern.test(`${provider} ${backend}`)) {
return { ok: false, reason: `provider/backend looks non-real: ${payload.provider}/${payload.backend}` };
}
return { ok: true, reason: "provider/model/trace/conversation evidence present" };
}
function summarizePayload(payload) {
const assistantReply = typeof payload?.reply?.content === "string" ? payload.reply.content.trim() : "";
const error = payload?.error
? {
code: payload.error.code,
missingEnv: payload.error.missingEnv,
missingCommands: payload.error.missingCommands,
providerStatus: payload.error.providerStatus
}
: null;
return {
status: payload?.status ?? null,
provider: payload?.provider ?? null,
model: payload?.model ?? null,
backend: payload?.backend ?? null,
assistantReplyNonEmpty: assistantReply.length > 0,
assistantReplyLength: assistantReply.length,
error
};
}
function parseArgs(argv) {
const parsed = { mode: "local", url: defaultLiveUrl, message: defaultLiveMessage };
for (let index = 0; index < argv.length; index += 1) {
const arg = argv[index];
if (arg === "--live") {
parsed.mode = "live";
} else if (arg === "--url") {
index += 1;
if (!argv[index]) throw new Error("--url requires a value");
parsed.url = argv[index];
} else if (arg === "--message") {
index += 1;
if (!argv[index]) throw new Error("--message requires a value");
parsed.message = argv[index];
} else if (arg === "--help") {
return { mode: "help" };
} else {
throw new Error(`unknown argument ${arg}`);
}
}
return parsed;
}
function usage() {
return {
status: "usage",
command: "node scripts/code-agent-chat-smoke.mjs [--live --url http://74.48.78.17:16666/]",
notes: [
"Default mode is local schema/readiness contract only and cannot pass #143 DEV-LIVE.",
"--live posts a minimal chat prompt to real DEV and passes only on completed plus non-empty assistant reply.",
"provider_unavailable with missing OPENAI_API_KEY or missing/forbidden HWLAB_CODE_AGENT_OPENAI_BASE_URL is reported as BLOCKED/provider-config."
]
};
}
function agentChatEndpoint(value) {
const url = new URL(value);
if (url.pathname.endsWith("/v1/agent/chat")) return url;
return new URL("/v1/agent/chat", url);
}
function postJson(url, { headers = {}, body, timeoutMs = 10000 }) {
const client = url.protocol === "https:" ? https : http;
const payload = JSON.stringify(body);
return new Promise((resolve, reject) => {
const request = client.request(
url,
{
method: "POST",
headers: {
"content-type": "application/json",
"content-length": Buffer.byteLength(payload),
...headers
},
insecureHTTPParser: true,
timeout: timeoutMs
},
(response) => {
const chunks = [];
response.on("data", (chunk) => chunks.push(chunk));
response.on("end", () => {
const text = Buffer.concat(chunks).toString("utf8");
resolve({
status: response.statusCode ?? null,
body: text,
json: parseJsonOrNull(text)
});
});
}
);
request.on("timeout", () => request.destroy(new Error(`request timed out after ${timeoutMs}ms`)));
request.on("error", reject);
request.end(payload);
});
}
function parseJsonOrNull(value) {
try {
return JSON.parse(value);
} catch {
return null;
}
}
function protocolId(prefix) {
return `${prefix}_${Date.now().toString(36)}`;
}