import {
MarksterToolDescriptor,
MarksterToolHttpMethod,
} from "../shared/marksterToolManifest";
import Ajv, { ValidateFunction } from "ajv";
import addFormats from "ajv-formats";
import fs from "node:fs";
import path from "node:path";
import { MARKSTER_TOOL_RESPONSE_SCHEMAS } from "./generated-openapi-schemas";
export type MarksterRuntimeMode = "live_with_fallback" | "live_strict" | "mock";
export type MarksterApiContext = {
baseUrl: string;
headers: Record<string, string>;
};
export class MarksterRequestError extends Error {
status: number;
body: unknown;
constructor(message: string, status: number, body: unknown) {
super(message);
this.name = "MarksterRequestError";
this.status = status;
this.body = body;
}
}
type MarksterToolCallInput = Record<string, unknown>;
type MockEnvelope = {
source: "mock";
tool: string;
method: MarksterToolHttpMethod;
endpoint: string;
path: string;
generated_at: string;
};
const DEFAULT_BASE_URL = "https://panel.markster.io";
const MOCK_INLINE_HEADSHOT_DIR = path.join(
process.cwd(),
"data/generated-headshots/inline-96",
);
const MOCK_HEADSHOT_BASE_URL = (() => {
const configuredHost =
process.env.MARKSTER_PUBLIC_BASE_URL ?? process.env.MCP_PUBLIC_BASE_URL;
if (configuredHost?.trim()) {
return `${configuredHost.replace(/\/$/, "")}/assets/headshots`;
}
// Default to deployed production asset host.
return "https://dealpulse-prod.hackathon.markster.io/assets/headshots";
})();
const MOCK_INLINE_HEADSHOT_DATA_URL_CACHE = new Map<string, string>();
const MOCK_PERSON_ID_BY_EMAIL: Record<string, string> = {
"dario@anthropic.com": "p1",
"sarah.chen@anthropic.com": "p2",
"marcus.j@anthropic.com": "p3",
"sam@openai.com": "p4",
"mina@openai.com": "p5",
"claire@stripe.com": "p6",
"diana@stripe.com": "p7",
"ivan@notion.so": "p8",
"daniel@ramp.com": "p9",
"nina@ramp.com": "p10",
"emily.hart@mercury.com": "p11",
"sasha@figma.com": "p12",
"reema@figma.com": "p13",
"maya@clay.com": "p14",
"pietro@manufact.com": "p15",
"luigi@manufact.com": "p16",
};
const ajv = new Ajv({
allErrors: true,
strict: false,
allowUnionTypes: true,
});
addFormats(ajv, ["date-time", "uuid", "email"]);
const TOOL_LOGGING_FLAG_KEYS = [
"MARKSTER_TOOL_CALL_LOGGING",
"MARKSTER_TOOL_LOGGING",
];
const DEFAULT_LIVE_TIMEOUT_MS = 12_000;
const DISABLED_TOOL_NAMES = new Set(["ai_team_memory_manage"]);
const MARKSTER_TOKEN_ENV_KEYS = [
"MARKSTER_CUSTOMER_API_KEY",
"MARKSTER_CUSTOMER_TOKEN",
"MARKSTER_CUSTOMER_API_TOKEN",
"MARKSTER_API_TOKEN",
"MARKSTER_API_KEY",
"MARKSTER_TOKEN",
];
let runtimeModeOverride: MarksterRuntimeMode | null = null;
function getTrimmedEnvValue(key: string): string | undefined {
const value = process.env[key];
if (value === undefined) return undefined;
const trimmed = value.trim();
return trimmed.length > 0 ? trimmed : undefined;
}
function nextRequestId(toolName: string): string {
return `${toolName}-${Date.now()}-${randomSuffix(String(Math.random()))}`;
}
const RESPONSE_VALIDATORS = new Map<string, ValidateFunction>();
const fallbackSchema = ajv.compile({
anyOf: [{ type: "object" }, { type: "array" }],
});
function getToolResponseValidator(toolName: string): ValidateFunction {
const existingValidator = RESPONSE_VALIDATORS.get(toolName);
if (existingValidator) {
return existingValidator;
}
const schema = (MARKSTER_TOOL_RESPONSE_SCHEMAS as Record<string, unknown>)[
toolName
];
const validate = schema
? ajv.compile(schema as Parameters<typeof ajv.compile>[0])
: fallbackSchema;
RESPONSE_VALIDATORS.set(toolName, validate);
return validate;
}
function validateMarksterPayload(toolName: string, payload: unknown): unknown {
const validate = getToolResponseValidator(toolName);
const valid = validate(payload);
if (!valid) {
throw new MarksterRequestError(
`Markster response did not match schema for tool '${toolName}'.`,
502,
{
tool: toolName,
issues: validate.errors,
sample: payload,
},
);
}
return payload;
}
function validateMockPayloadBestEffort(
toolName: string,
payload: unknown,
): unknown {
try {
return validateMarksterPayload(toolName, payload);
} catch (error) {
const reason = error instanceof Error ? error.message : String(error);
logToolEvent(
"error",
`mock_schema_mismatch name=${toolName} reason=${reason} (continuing with mock payload)`,
);
return payload;
}
}
function isObject(value: unknown): value is Record<string, unknown> {
return typeof value === "object" && value !== null && !Array.isArray(value);
}
function asRecord(value: unknown): MarksterToolCallInput {
return isObject(value) ? { ...value } : {};
}
function boolFromEnv(value?: string): boolean {
if (!value) return false;
return ["1", "true", "yes", "on", "enabled"].includes(value.toLowerCase());
}
function parseRuntimeMode(value?: string): MarksterRuntimeMode | undefined {
if (!value) return undefined;
const normalized = value.trim().toLowerCase();
if (
normalized === "live_with_fallback" ||
normalized === "live-fallback" ||
normalized === "live"
) {
return "live_with_fallback";
}
if (
normalized === "live_strict" ||
normalized === "live-strict" ||
normalized === "strict"
) {
return "live_strict";
}
if (normalized === "mock" || normalized === "demo") {
return "mock";
}
return undefined;
}
function resolveRuntimeModeFromEnv(): MarksterRuntimeMode {
const explicitRuntime = parseRuntimeMode(
getTrimmedEnvValue("MARKSTER_RUNTIME_MODE"),
);
if (explicitRuntime) {
return explicitRuntime;
}
if (boolFromEnv(process.env.MARKSTER_MOCK_MODE)) {
return "mock";
}
return "mock";
}
export function getMarksterRuntimeMode(): MarksterRuntimeMode {
return runtimeModeOverride ?? resolveRuntimeModeFromEnv();
}
export function setMarksterRuntimeMode(
mode: MarksterRuntimeMode,
): MarksterRuntimeMode {
runtimeModeOverride = mode;
return getMarksterRuntimeMode();
}
export function clearMarksterRuntimeModeOverride(): void {
runtimeModeOverride = null;
}
export function shouldLogToolCalls(): boolean {
for (const key of TOOL_LOGGING_FLAG_KEYS) {
const envValue = process.env[key];
if (envValue !== undefined) {
return boolFromEnv(envValue);
}
}
return true;
}
function logToolEvent(level: "debug" | "error", message: string): void {
if (!shouldLogToolCalls()) return;
if (level === "error") {
console.error(`[markster:tool] ${message}`);
return;
}
console.info(`[markster:tool] ${message}`);
}
function nowISO(): string {
return new Date().toISOString();
}
function randomSuffix(seed: string): string {
let hash = 0;
for (let i = 0; i < seed.length; i += 1) {
hash = (hash * 31 + seed.charCodeAt(i)) % 1000000007;
}
return String(hash).padStart(6, "0");
}
function normalizeEmailKey(value: string | undefined): string | undefined {
if (!value) return undefined;
const trimmed = value.trim().toLowerCase();
return trimmed.length > 0 ? trimmed : undefined;
}
function mockHeadshotUrlForPersonId(
personId: string | undefined,
): string | undefined {
if (!personId) return undefined;
const normalized = personId.trim().toLowerCase();
if (!/^p\d+$/.test(normalized)) return undefined;
const cachedDataUrl = MOCK_INLINE_HEADSHOT_DATA_URL_CACHE.get(normalized);
if (cachedDataUrl) {
return cachedDataUrl;
}
const inlinePath = path.join(MOCK_INLINE_HEADSHOT_DIR, `${normalized}.jpg`);
try {
const buffer = fs.readFileSync(inlinePath);
if (buffer.length > 0) {
const dataUrl = `data:image/jpeg;base64,${buffer.toString("base64")}`;
MOCK_INLINE_HEADSHOT_DATA_URL_CACHE.set(normalized, dataUrl);
return dataUrl;
}
} catch {
// Fall through to URL path fallback for environments without inline assets.
}
return `${MOCK_HEADSHOT_BASE_URL}/${normalized}.png`;
}
function resolveMockPersonIdForEmail(
email: string | undefined,
): string | undefined {
const normalizedEmail = normalizeEmailKey(email);
if (!normalizedEmail) return undefined;
return MOCK_PERSON_ID_BY_EMAIL[normalizedEmail];
}
function getMarksterTokenFromEnv(): string | undefined {
for (const key of MARKSTER_TOKEN_ENV_KEYS) {
const value = getTrimmedEnvValue(key);
if (value) {
return value;
}
}
return undefined;
}
function hasAnyMarksterEnvValue(): boolean {
return MARKSTER_TOKEN_ENV_KEYS.some(
(key) => getTrimmedEnvValue(key) !== undefined,
);
}
function tokenEnvPresenceSummary(): string {
return MARKSTER_TOKEN_ENV_KEYS.map((key) => {
const value = getTrimmedEnvValue(key);
return `${key}=${value ? "set" : "unset"}`;
}).join(", ");
}
export function createMarksterContext(): MarksterApiContext {
const baseUrl =
process.env.MARKSTER_API_BASE_URL ??
process.env.MARKSTER_PANEL_URL ??
process.env.MARKSTER_API_URL ??
DEFAULT_BASE_URL;
const token = getMarksterTokenFromEnv();
if (!token) {
throw new Error(
"Missing Markster customer API token. Set MARKSTER_CUSTOMER_API_KEY (preferred, or MARKSTER_API_TOKEN fallback) and keep MARKSTER_MOCK_MODE=false/unset.",
);
}
const headers: Record<string, string> = {
Accept: "application/json",
"Content-Type": "application/json",
"User-Agent": "markster-business-os/0.2.0",
};
if (token) {
headers.Authorization = `Bearer ${token}`;
headers["X-API-Key"] = token;
}
return { baseUrl: baseUrl.replace(/\/$/, ""), headers };
}
export function shouldUseMockMode(): boolean {
const runtimeMode = getMarksterRuntimeMode();
if (runtimeMode === "mock") {
return true;
}
if (!getMarksterTokenFromEnv()) {
if (hasAnyMarksterEnvValue()) {
console.warn(
`[markster] No usable Markster token found. Checked env: ${tokenEnvPresenceSummary()}`,
);
} else {
console.warn(
"[markster] No Markster API token env vars set; falling back to MOCK mode.",
);
}
return true;
}
return false;
}
type InterpolationResult = {
path: string;
missing: string[];
remainingArgs: MarksterToolCallInput;
};
function interpolateEndpoint(
endpoint: string,
args: MarksterToolCallInput,
): InterpolationResult {
const missing: string[] = [];
const remaining = { ...args };
const path = endpoint.replace(/\{([a-zA-Z0-9_-]+)\}/g, (_match, key) => {
const value = remaining[key];
if (value === undefined || value === null) {
missing.push(key);
return `:${key}`;
}
delete remaining[key];
return encodeURIComponent(String(value));
});
return { path, missing, remainingArgs: remaining };
}
export async function callMarksterTool(
spec: MarksterToolDescriptor,
args: unknown,
): Promise<unknown> {
if (DISABLED_TOOL_NAMES.has(spec.name)) {
throw new Error(`Tool '${spec.name}' is disabled for this MCP demo.`);
}
const requestArgs: MarksterToolCallInput = asRecord(args);
const interpolation = interpolateEndpoint(spec.endpoint, requestArgs);
const requestId = nextRequestId(spec.name);
if (interpolation.missing.length > 0) {
logToolEvent(
"error",
`request=${requestId} missing path params for ${
spec.name
}: ${interpolation.missing.join(", ")}`,
);
throw new Error(
`Missing required path parameter(s): ${interpolation.missing.join(
", ",
)} for ${spec.name}.`,
);
}
const cleanedArgs = interpolation.remainingArgs;
const query = Object.keys(cleanedArgs).length
? JSON.stringify(cleanedArgs)
: "{}";
const startedAt = Date.now();
logToolEvent(
"debug",
`request=${requestId} start name=${spec.name} method=${spec.method} path=${interpolation.path} query=${query}`,
);
const runtimeMode = getMarksterRuntimeMode();
const fallbackEnabled = runtimeMode === "live_with_fallback";
const mockOnlyMode = runtimeMode === "mock";
if (mockOnlyMode) {
const mockPayload = buildMockPayload(spec, interpolation.path, cleanedArgs);
const validatedPayload = validateMockPayloadBestEffort(
spec.name,
mockPayload.payload,
);
const elapsedMs = Date.now() - startedAt;
logToolEvent(
"debug",
`request=${requestId} done name=${spec.name} status=mock_ok duration_ms=${elapsedMs}`,
);
return {
status: 200,
ok: true,
runtime_mode: runtimeMode,
...mockPayload,
payload: validatedPayload,
};
}
try {
const context = createMarksterContext();
const [path, queryString] = splitPathAndQuery(interpolation.path);
const { explicitQuery, requestBody } = extractBodyAndQuery(cleanedArgs);
const requestPath = buildRequestUrl(
path,
queryString,
explicitQuery,
cleanedArgs,
context.baseUrl,
spec.method,
);
const requestInit = buildRequestInit(
spec.method,
context.headers,
requestBody,
cleanedArgs,
);
const timeoutMs = resolveLiveTimeoutMs();
const abortController = new AbortController();
const timeoutHandle = setTimeout(() => {
abortController.abort();
}, timeoutMs);
let response: Response;
try {
response = await fetch(requestPath, {
...requestInit,
signal: abortController.signal,
});
} catch (error) {
if (abortController.signal.aborted) {
throw new Error(
`Markster live request timed out after ${timeoutMs}ms for ${spec.name}.`,
);
}
throw error;
} finally {
clearTimeout(timeoutHandle);
}
const elapsedMs = Date.now() - startedAt;
const text = await response.text();
let payload: unknown = null;
if (text) {
try {
payload = JSON.parse(text);
} catch {
payload = text;
}
}
if (!response.ok) {
logToolEvent(
"error",
`request=${requestId} failed name=${spec.name} status=${response.status} endpoint=${requestPath}`,
);
throw new MarksterRequestError(
`Markster API call failed: ${spec.name}`,
response.status,
payload,
);
}
const validatedPayload = validateMarksterPayload(spec.name, payload);
logToolEvent(
"debug",
`request=${requestId} done name=${spec.name} status=${response.status} duration_ms=${elapsedMs}`,
);
return {
status: response.status,
ok: true,
tool: spec.name,
method: spec.method,
endpoint: spec.endpoint,
runtime_mode: runtimeMode,
payload: validatedPayload,
};
} catch (error) {
if (!fallbackEnabled || !shouldFallbackToMock(error)) {
throw error;
}
const fallbackMeta = formatFallbackMeta(error);
logToolEvent(
"error",
`request=${requestId} fallback_to_mock name=${spec.name} reason=${fallbackMeta.message}`,
);
const mockPayload = buildMockPayload(spec, interpolation.path, cleanedArgs);
const validatedPayload = validateMockPayloadBestEffort(
spec.name,
mockPayload.payload,
);
const elapsedMs = Date.now() - startedAt;
logToolEvent(
"debug",
`request=${requestId} done name=${spec.name} status=mock_fallback duration_ms=${elapsedMs}`,
);
return {
status: 200,
ok: true,
runtime_mode: runtimeMode,
fallback: {
active: true,
reason: fallbackMeta.message,
error_type: fallbackMeta.type,
...(fallbackMeta.status !== undefined
? { status: fallbackMeta.status }
: {}),
},
...mockPayload,
payload: validatedPayload,
};
}
}
function resolveLiveTimeoutMs(): number {
const fromEnv = Number.parseInt(
process.env.MARKSTER_API_TIMEOUT_MS ?? "",
10,
);
if (Number.isFinite(fromEnv) && fromEnv > 0) {
return fromEnv;
}
return DEFAULT_LIVE_TIMEOUT_MS;
}
function shouldFallbackToMock(error: unknown): boolean {
if (error instanceof MarksterRequestError) {
return (
error.status === 401 ||
error.status === 403 ||
error.status === 408 ||
error.status === 429 ||
error.status >= 500
);
}
return true;
}
function formatFallbackMeta(error: unknown): {
type: string;
message: string;
status?: number;
} {
if (error instanceof MarksterRequestError) {
return {
type: "markster_request_error",
message: `${error.message} (${error.status})`,
status: error.status,
};
}
if (error instanceof Error) {
return {
type: "error",
message: error.message,
};
}
return {
type: "unknown",
message: String(error),
};
}
function extractBodyAndQuery(args: MarksterToolCallInput): {
explicitQuery?: Record<string, unknown>;
requestBody?: Record<string, unknown>;
} {
const body = isObject(args.body) ? args.body : undefined;
const query = isObject(args.query) ? args.query : undefined;
return { explicitQuery: query, requestBody: body };
}
function splitPathAndQuery(pathWithQuery: string): [string, URLSearchParams] {
const [pathPart, queryPart] = pathWithQuery.split("?", 2);
const query = new URLSearchParams(queryPart);
return [pathPart, query];
}
function buildRequestUrl(
basePath: string,
existingQuery: URLSearchParams,
explicitQuery: Record<string, unknown> | undefined,
requestArgs: MarksterToolCallInput,
baseUrl: string,
method: MarksterToolHttpMethod,
): string {
if (method !== "GET") {
return contextualUrl(basePath, existingQuery, baseUrl);
}
const additionalQuery = new URLSearchParams(existingQuery);
addRequestArgsToQuery(additionalQuery, requestArgs);
addObjectToQuery(additionalQuery, explicitQuery);
return contextualUrl(basePath, additionalQuery, baseUrl);
}
function addRequestArgsToQuery(
query: URLSearchParams,
requestArgs: MarksterToolCallInput,
): void {
Object.entries(requestArgs).forEach(([key, value]) => {
if (key === "body" || key === "query") {
return;
}
if (value === undefined || value === null) {
return;
}
if (Array.isArray(value)) {
value.forEach((item) => {
if (item !== undefined && item !== null)
query.append(key, String(item));
});
return;
}
if (isObject(value)) {
return;
}
query.set(key, String(value));
});
}
function addObjectToQuery(
query: URLSearchParams,
objectValue?: Record<string, unknown>,
): void {
if (!isObject(objectValue)) return;
Object.entries(objectValue).forEach(([key, value]) => {
if (value === undefined || value === null) return;
if (Array.isArray(value) || isObject(value)) return;
query.set(key, String(value));
});
}
function buildRequestInit(
method: MarksterToolHttpMethod,
headers: Record<string, string>,
requestBody: Record<string, unknown> | undefined,
requestArgs: MarksterToolCallInput,
): RequestInit {
const requestInit: RequestInit = {
method,
headers,
};
if (method === "GET" || method === "DELETE") {
return requestInit;
}
let bodyPayload: Record<string, unknown> | undefined = requestBody;
if (bodyPayload === undefined) {
const inlineBody = { ...requestArgs };
delete inlineBody.body;
delete inlineBody.query;
bodyPayload = Object.keys(inlineBody).length > 0 ? inlineBody : undefined;
}
if (bodyPayload !== undefined) {
requestInit.body = JSON.stringify(bodyPayload);
}
return requestInit;
}
function contextualUrl(
path: string,
query: URLSearchParams,
baseUrl: string,
): string {
const hasQuery = query.toString();
return `${baseUrl}${path}${hasQuery ? `?${hasQuery}` : ""}`;
}
function buildMockPayload(
spec: MarksterToolDescriptor,
path: string,
args: MarksterToolCallInput,
): MockEnvelope & { payload: unknown; tool: string } {
const seed = `${spec.name}:${path}:${JSON.stringify(args)}`;
const generatedAt = nowISO();
const base = {
source: "mock" as const,
tool: spec.name,
method: spec.method,
endpoint: spec.endpoint,
path,
generated_at: generatedAt,
};
const requestId = `mock-${randomSuffix(seed)}`;
const sampleTaskId = String(args.task_id ?? "task-101");
const sampleConversationId = String(args.conversation_id ?? "conv-88");
const sampleCalendarId = String(args.calendar_id ?? "calendar-77");
const sampleItemId = String(args.item_id ?? "post-33");
const sampleSequenceId = String(args.sequence_id ?? "seq-09");
const sampleLeadIds = Array.isArray(args.lead_ids)
? args.lead_ids.filter((item): item is string => typeof item === "string")
: ["lead-1", "lead-2"];
switch (spec.name) {
case "customer_workspace_snapshot": {
return {
...base,
payload: {
plan_tier: "growth",
calendar_link: "https://cal.com/markster/demo",
mode: "live",
custom_fields: {
operating_model: "ai_native",
product: "markster_business_os",
},
whatsapp_morning_brief_enabled: true,
id: 13,
company_id: 101,
onboarding_state: {
current_step: 7,
total_steps: 7,
completion_pct: 100,
},
onboarding_completed: 1,
preferred_language: "en",
whatsapp_morning_brief_last_sent_at: generatedAt,
created_at: "2025-09-01T09:00:00Z",
updated_at: generatedAt,
company: {
name: "Markster",
website: "https://markster.ai",
business_type: "professional_services",
industry: "Revenue operations",
google_folder_id: null,
replyio_workspace_id: null,
usp: "AI-native GTM operating system",
icp: "Founder-led B2B service companies",
business_goals: "pipeline growth, predictable meetings",
industry_trends_email: null,
id: 101,
created_at: "2025-09-01T09:00:00Z",
updated_at: generatedAt,
onboarding_profile: {
timezone: "America/Los_Angeles",
channels: ["email", "linkedin"],
},
primary_offer: "Growth",
target_role: "Founder / VP Revenue",
buyer_geo: "US",
availability_days: "Mon-Fri",
availability_hours_start: "09:00",
availability_hours_end: "17:00",
meeting_duration_min: 30,
meeting_modes: "remote",
office_address: null,
on_site_radius_km: null,
on_site_fee_note: null,
weekly_bookings_goal: 12,
avg_deal_value: 50000,
allowed_channels: "email,linkedin",
timezone: "America/Los_Angeles",
},
contacts: [
{
first_name: "Ivan",
last_name: "Ivanka",
email: "ivan@markster.ai",
phone: "+1-415-555-0101",
position: "Founder",
is_primary: 1,
id: 501,
company_id: 101,
created_at: "2025-09-01T09:00:00Z",
updated_at: generatedAt,
},
],
profile: {
id: "cust_mrkstr",
company: "Markster",
plan: "Growth",
owner: "Ivan Ivanka",
},
notifications: [
{
id: "n-1",
type: "outreach",
title:
"Figma: Reema Batta visited pricing 6x — intent 87, ready to buy",
status: "success",
},
{
id: "n-2",
type: "outreach",
title: "Anthropic: Sarah Chen submitted demo request",
status: "success",
},
{
id: "n-3",
type: "task",
title: "Kontext Group quarterly review due tomorrow",
status: "info",
},
],
onboarding: {
current_step: 7,
total_steps: 7,
completion_pct: 100,
started_at: "2026-01-15T08:00:00Z",
},
tasks: {
open_count: 8,
completed_count: 23,
blocked_count: 1,
overdue_count: 0,
},
cold_mail_setup: {
status: "ready",
domain_health: "good",
warmup_day: 21,
warmup_target_days: 21,
},
dashboard: {
kpis: [
{ key: "mrr", value: "$15,628", trend: "+18.3% MoM" },
{ key: "pipeline", value: "$1.25M", trend: "+42% MoM" },
{
key: "total_revenue",
value: "$211,534",
trend: "+$24K one-off",
},
{ key: "gross_margin", value: "82%", trend: "+4.2pp" },
{
key: "meetings_booked",
value: "8 this month",
trend: "+33% MoM",
},
{ key: "win_rate", value: "34%", trend: "+6pp QoQ" },
],
},
requestId,
},
};
}
case "customer_profile_manage": {
const update: Record<string, unknown> | undefined = isObject(args.updates)
? args.updates
: isObject(args.profile)
? args.profile
: undefined;
const action =
typeof args.action === "string"
? args.action
: typeof args.mode === "string"
? args.mode
: "read";
return {
...base,
payload: {
plan_tier: "growth",
calendar_link: "https://cal.com/markster/demo",
mode: "live",
custom_fields: {
operating_model: "ai_native",
},
whatsapp_morning_brief_enabled: true,
id: 13,
company_id: 101,
onboarding_state: {
current_step: 7,
total_steps: 7,
completion_pct: 100,
},
onboarding_completed: 1,
preferred_language: "en",
whatsapp_morning_brief_last_sent_at: generatedAt,
created_at: "2025-09-01T09:00:00Z",
updated_at: generatedAt,
company: {
name: "Markster",
website: "https://markster.ai",
business_type: "professional_services",
industry: "Revenue operations",
google_folder_id: null,
replyio_workspace_id: null,
usp: "AI-native GTM operating system",
icp: "Founder-led B2B service companies",
business_goals: "pipeline growth, predictable meetings",
industry_trends_email: null,
id: 101,
created_at: "2025-09-01T09:00:00Z",
updated_at: generatedAt,
onboarding_profile: {
timezone: "America/Los_Angeles",
channels: ["email", "linkedin"],
},
primary_offer: "Growth",
target_role: "Founder / VP Revenue",
buyer_geo: "US",
availability_days: "Mon-Fri",
availability_hours_start: "09:00",
availability_hours_end: "17:00",
meeting_duration_min: 30,
meeting_modes: "remote",
office_address: null,
on_site_radius_km: null,
on_site_fee_note: null,
weekly_bookings_goal: 12,
avg_deal_value: 50000,
allowed_channels: "email,linkedin",
timezone: "America/Los_Angeles",
},
contacts: [
{
first_name: "Ivan",
last_name: "Ivanka",
email: "ivan@markster.ai",
phone: "+1-415-555-0101",
position: "Founder",
is_primary: 1,
id: 501,
company_id: 101,
created_at: "2025-09-01T09:00:00Z",
updated_at: generatedAt,
},
],
action,
profile_update: update,
requestId,
},
};
}
case "contact_profile_workspace":
case "contact_profile_workspace_get": {
const includeEnriched = args.include_enriched !== false;
const filterContactId =
typeof args.contact_id === "string" ? args.contact_id : "c1";
const filterEmail = typeof args.email === "string" ? args.email : null;
const mkContact = (
id: string,
name: string,
email: string,
title: string,
company: string,
linkedin: string,
city: string,
intentScore: number,
intentStatus: string,
) => {
const personId = resolveMockPersonIdForEmail(email);
const photoUrl = mockHeadshotUrlForPersonId(personId);
return {
contact_id: id,
person_id: personId,
name,
email,
title,
company,
source: "crm",
intent_score: intentScore,
intent_status: intentStatus,
...(photoUrl ? { photo_url: photoUrl, avatar_url: photoUrl } : {}),
enriched: includeEnriched
? {
contactdb: {
matched: true,
match_score: 0.94,
data: {
id: `person-${id}`,
person_id: personId,
full_name: name,
title,
company_name: company,
email,
linkedin_url: linkedin,
city,
country: "United States",
...(photoUrl
? { photo_url: photoUrl, avatar_url: photoUrl }
: {}),
},
},
}
: {
contactdb: {
matched: false,
match_score: 0,
data: null,
},
},
};
};
const allContacts = [
mkContact(
"c1",
"Claire Hughes",
"claire@stripe.com",
"VP Revenue",
"Stripe",
"https://www.linkedin.com/in/claire-hughes",
"San Francisco",
89,
"Ready to Buy",
),
mkContact(
"c2",
"Sarah Chen",
"sarah.chen@anthropic.com",
"VP Revenue",
"Anthropic",
"https://www.linkedin.com/in/sarah-chen-ai",
"San Francisco",
88,
"Ready to Buy",
),
mkContact(
"c3",
"Maya Chen",
"maya@clay.com",
"Chief Revenue Officer",
"Clay",
"https://www.linkedin.com/in/maya-chen-cro",
"New York",
87,
"Ready to Buy",
),
mkContact(
"c4",
"Emily Hart",
"emily.hart@mercury.com",
"VP Sales",
"Mercury",
"https://www.linkedin.com/in/emily-hart-sales",
"San Francisco",
84,
"Ready to Buy",
),
mkContact(
"c5",
"Mina Patel",
"mina@openai.com",
"Director of Partnerships",
"OpenAI",
"https://www.linkedin.com/in/mina-patel",
"San Francisco",
82,
"Ready to Buy",
),
mkContact(
"c6",
"Luigi Pignone",
"luigi@manufact.com",
"Co-Founder",
"Manufact",
"https://www.linkedin.com/in/luigi-pignone",
"San Francisco",
81,
"Ready to Buy",
),
mkContact(
"c7",
"Dario Amodei",
"dario@anthropic.com",
"CEO",
"Anthropic",
"https://www.linkedin.com/in/dario",
"San Francisco",
76,
"Hot",
),
mkContact(
"c8",
"Pietro Zullo",
"pietro@manufact.com",
"Founder",
"Manufact",
"https://www.linkedin.com/in/pietro-zullo",
"San Francisco",
74,
"Hot",
),
mkContact(
"c9",
"Diana Lin",
"diana@stripe.com",
"Head of Enterprise",
"Stripe",
"https://www.linkedin.com/in/diana-lin",
"San Francisco",
73,
"Hot",
),
mkContact(
"c10",
"Daniel Brooks",
"daniel@ramp.com",
"Chief Revenue Officer",
"Ramp",
"https://www.linkedin.com/in/daniel-brooks-cro",
"New York",
70,
"Hot",
),
mkContact(
"c11",
"Ivan Ivanov",
"ivan@notion.so",
"CRO",
"Notion",
"https://www.linkedin.com/in/ivan-ivanov-notion",
"San Francisco",
69,
"Hot",
),
mkContact(
"c12",
"Sasha Ivanov",
"sasha@figma.com",
"VP Revenue",
"Figma",
"https://www.linkedin.com/in/sasha-ivanov-figma",
"San Francisco",
68,
"Hot",
),
mkContact(
"c13",
"Sam Altman",
"sam@openai.com",
"Chief Executive Officer",
"OpenAI",
"https://www.linkedin.com/in/sam-altman",
"San Francisco",
57,
"Warming",
),
mkContact(
"c14",
"Nina K",
"nina@ramp.com",
"Head of Mid-Market",
"Ramp",
"https://www.linkedin.com/in/nina-k-ramp",
"New York",
53,
"Warming",
),
mkContact(
"c15",
"Marcus Johnson",
"marcus.j@anthropic.com",
"Head of Sales",
"Anthropic",
"https://www.linkedin.com/in/marcus-johnson-sales",
"San Francisco",
49,
"Warming",
),
mkContact(
"c16",
"Reema Batta",
"reema@figma.com",
"VP Growth Marketing",
"Figma",
"https://www.linkedin.com/in/reema-batta/",
"San Francisco",
46,
"Warming",
),
];
const filtered = allContacts.filter((contact) => {
if (
typeof args.contact_id === "string" &&
contact.contact_id !== args.contact_id
) {
return false;
}
if (filterEmail && contact.email !== filterEmail) {
return false;
}
if (
typeof args.name === "string" &&
!contact.name.toLowerCase().includes(args.name.toLowerCase())
) {
return false;
}
if (
typeof args.title === "string" &&
!contact.title.toLowerCase().includes(args.title.toLowerCase())
) {
return false;
}
if (
typeof args.company === "string" &&
!contact.company.toLowerCase().includes(args.company.toLowerCase())
) {
return false;
}
return true;
});
if (spec.name === "contact_profile_workspace_get") {
return {
...base,
payload:
filtered[0] ??
mkContact(
filterContactId,
"Unknown Contact",
"unknown@example.com",
"Unknown",
"Unknown",
"",
"",
0,
"Unknown",
),
};
}
return {
...base,
payload: {
total: filtered.length,
include_enriched: includeEnriched,
filters: {
contact_id:
typeof args.contact_id === "string" ? args.contact_id : undefined,
name: typeof args.name === "string" ? args.name : undefined,
email: filterEmail ?? undefined,
title: typeof args.title === "string" ? args.title : undefined,
company:
typeof args.company === "string" ? args.company : undefined,
source: typeof args.source === "string" ? args.source : "all",
},
results: filtered,
requestId,
},
};
}
case "onboarding_flow_manage": {
const responses = isObject(args.step_answers)
? args.step_answers
: isObject(args.answers)
? args.answers
: null;
return {
...base,
payload: {
id: 9001,
survey_id: 77,
customer_id: 13,
responses,
current_step: 5,
is_completed: 0,
started_at: "2026-02-20T09:00:00Z",
completed_at: null,
flow_id: `flow-${randomSuffix(
String(args.company_key ?? "default"),
)}`,
status: "ok",
stage: args.stage || "assessment",
answers_stored: responses ? Object.keys(responses).length : 0,
requestId,
},
};
}
case "onboarding_ai_assist": {
return {
...base,
payload: {
suggestions: [
"Suggested ICP title: Agency Owner / VP Revenue / Head of Growth",
"Suggested voice: practical, concise, founder-first — lead with results not theory",
],
quality_score: 91,
generated_profile_fields: {
risk_profile: "balanced",
onboarding_complexity: "medium",
},
requestId,
},
};
}
case "dashboard_insights": {
return {
...base,
payload: {
customer_id: 13,
period: typeof args.period === "string" ? args.period : "daily",
qualified_meetings: 8,
show_rate: 78.5,
positive_reply_rate: 6.8,
bookings_per_1k_recipients: 4.1,
deliverability_health_index: 92.4,
hard_bounce_rate: 0.7,
complaint_rate: 0.12,
messages_sent: 1240,
leads_generated: 37,
daily_send_pace: 177,
sql_rate: 22.4,
time_to_first_send_hours: 6.2,
time_to_first_meeting_hours: 51.4,
revenue: 15627.83,
previous_qualified_meetings: 6,
previous_messages_sent: 1110,
previous_revenue: 13205.17,
previous_positive_reply_rate: 5.9,
previous_show_rate: 73.2,
previous_bookings_per_1k_recipients: 3.6,
refresh: true,
time_series: [
{ date: "2025-11-01", mrr: 9811.0, leads: 12 },
{ date: "2025-12-01", mrr: 11043.5, leads: 16 },
{ date: "2026-01-01", mrr: 13205.17, leads: 22 },
{ date: "2026-02-01", mrr: 15627.83, leads: 34 },
],
stages: [
{ name: "Prospect", count: 10, value: "$545,000" },
{ name: "Discovery", count: 5, value: "$295,000" },
{ name: "Proposal", count: 3, value: "$185,000" },
{ name: "Negotiation", count: 2, value: "$222,500" },
],
rollup: {
total_revenue: "$187,534 ARR",
pipeline_value: "$1,247,500",
win_rate: "34%",
avg_deal_size: "$103,958",
open_opportunities: 12,
active_contacts: 22,
meetings_booked_this_month: 5,
team_velocity: "+22.3%",
},
requestId,
},
};
}
case "task_workspace": {
return {
...base,
payload: {
workspace: "sales",
overview: {
active: 8,
completed_today: 4,
stalled: 1,
overdue: 0,
},
tasks: [
{
id: "task-101",
title:
"Close Figma deal — Reema Batta pricing review complete (intent 87)",
status: "in_progress",
owner: "ivan",
priority: "urgent",
due: "2026-02-21",
type: "deal",
context:
"Reema visited pricing 6x, used ROI calculator, and reviewed the case study. $150K deal in negotiation. Close probability: 72%.",
},
{
id: "task-106",
title:
"Reply to Reema Batta — follow-up email on competitive positioning",
status: "queued",
owner: "ivan",
priority: "urgent",
due: "2026-02-21",
type: "email",
context:
"Reema replied asking how Markster compares to traditional agencies. Draft response using founder voice — lead with 4 agencies, 12 years, zero employees proof point.",
},
{
id: "task-102",
title: "Send Anthropic enterprise proposal to Sarah Chen",
status: "queued",
owner: "ivan",
priority: "high",
due: "2026-02-22",
type: "deal",
context:
"Sarah used ROI calculator twice and downloaded 3 case studies. Deal size: $120K.",
},
{
id: "task-107",
title:
"Reply to Datadog intro — Alex Kim wants technical deep-dive",
status: "queued",
owner: "ivan",
priority: "high",
due: "2026-02-22",
type: "email",
context:
"Alex forwarded our case study to his VP. Wants a technical walkthrough of the platform. Reply with available slots this week.",
},
{
id: "task-103",
title: "Schedule pricing walkthrough with Clay — Maya Chen",
status: "queued",
owner: "ivan",
priority: "high",
due: "2026-02-23",
type: "deal",
context:
"Maya submitted demo request. As CRO, she's the decision maker. Intent score: 87.",
},
{
id: "task-105",
title:
"Send Figma competitive analysis — Reema requested comparison",
status: "queued",
owner: "ivan",
priority: "high",
due: "2026-02-22",
type: "content",
context:
"Reema asked for competitive comparison with traditional agencies. Send our positioning doc.",
},
{
id: "task-104",
title: "Publish case study — 9.2x revenue growth in 15 months",
status: "queued",
owner: "automation",
priority: "normal",
due: "2026-02-24",
type: "content",
context:
"LinkedIn + blog distribution. Content approved, awaiting publish slot.",
},
],
requestId,
},
};
}
case "task_action": {
return {
...base,
payload: {
task_id: sampleTaskId,
action: args.action || "start",
previous_status: "queued",
status: args.action === "complete" ? "complete" : "in_progress",
note: "Mock task transition saved.",
requestId,
},
};
}
case "task_content_regenerate": {
return {
...base,
payload: {
task_id: sampleTaskId,
regenerate_token: `regen-${randomSuffix(sampleTaskId)}`,
status: "ok",
revised_content: `Updated onboarding content block for ${sampleTaskId}`,
requestId,
},
};
}
case "ai_team_workspace": {
return {
...base,
payload: {
agents: [
{ id: "agent-research", name: "Researcher", role: "insights" },
{ id: "agent-finance", name: "Finance Copilot", role: "modeling" },
{ id: "agent-copy", name: "Positioning Writer", role: "content" },
],
conversations: [
{
id: sampleConversationId,
name: "Stripe enterprise deal strategy",
updated: generatedAt,
},
],
requestId,
},
};
}
case "ai_team_chat_turn": {
return {
...base,
payload: {
conversation_id: sampleConversationId,
turn: 3,
user: {
text:
typeof args.message === "string"
? args.message
: "Mock user turn",
},
assistant:
"I analyzed Markster's pipeline and recommend prioritizing Stripe — Claire Hughes has visited pricing 5x this week, used the ROI calculator, and submitted a demo request. Intent score: 89.",
sse_simulated: true,
requestId,
},
};
}
case "ai_team_memory_manage": {
const operation =
typeof args.operation === "string" ? args.operation : "list";
return {
...base,
payload: {
operation,
memory_items: [
{
id: "mem-001",
label: "Stripe: Claire Hughes opened pricing page 5x this week",
pinned: true,
},
{
id: "mem-002",
label:
"Kontext Group prefers results-first messaging — always lead with revenue numbers",
pinned: false,
},
],
requestId,
},
};
}
case "customer_artifact_workspace": {
const operation =
typeof args.operation === "string" ? args.operation : "list";
return {
...base,
payload: {
operation,
artifacts: [
{
id: "art-9",
title: "Markster GTM Playbook",
status: "published",
updated_at: generatedAt,
},
],
requestId,
},
};
}
case "customer_document_workspace": {
return {
...base,
payload: {
documents: [
{
id: "doc-1",
title: "Q1 GTM Strategy — Agency ICP Expansion",
kind: "md",
},
{
id: "doc-2",
title: "Kontext Group Case Study — $71K to $1.3M",
kind: "txt",
},
],
requestId,
},
};
}
case "customer_data_change_review": {
return {
...base,
payload: {
proposals: [
{
id: "prop-12",
area: "voice_pack",
change: "Tone updated from pragmatic to pragmatic+ambitious",
status: "pending",
requested_by: "onboarding_ai_assist",
},
],
requestId,
},
};
}
case "content_calendar_workspace": {
// Raw endpoint shape for GET /api/content/calendars:
// Always return a list of calendars. Curated week filtering is handled by
// runContentCalendarWorkspace in src/tools/markster-business-os.ts.
return {
...base,
payload: [
{
id: sampleCalendarId,
name: "Current content calendar",
status: "ready",
date_start: "2026-02-16",
date_end: "2026-02-22",
items: 14,
},
{
id: "41",
name: "Previous cycle",
status: "ready",
date_start: "2026-02-09",
date_end: "2026-02-15",
items: 11,
},
],
};
}
case "content_calendar_workspace_get": {
const calendarId =
typeof args.calendar_id === "string" ? args.calendar_id : "42";
return {
...base,
payload: {
id: calendarId,
name: "Current content calendar",
status: "ready",
date_start: "2026-02-16",
date_end: "2026-02-22",
items: [
{
id: "501",
date: "2026-02-20",
platform: "LinkedIn",
pillar: "Case Study",
daily_topic:
"We grew Kontext Group from $71K to $1.3M. Here's what moved the needle.",
status: "scheduled",
generated_post: {
id: "9001",
title:
"We grew Kontext Group from $71K to $1.3M. Here's what moved the needle.",
hook: "Everyone talks about AI replacing agencies. We used it to 18x a client's revenue.",
body: "In 2023, Kontext Group was doing $71K. Two years later, they crossed $1.3M...",
image_url:
"https://dealpulse-prod.hackathon.markster.io/assets/headshots/p2.png",
status: "draft",
},
},
{
id: "502",
date: "2026-02-21",
platform: "LinkedIn",
pillar: "Owner Independence",
daily_topic:
"Why your agency doesn't need 50 employees to hit $2M ARR",
status: "scheduled",
generated_post: {
id: "9002",
title:
"Why your agency doesn't need 50 employees to hit $2M ARR",
hook: "Most growth teams scale headcount first. We scaled systems first.",
body: "The old playbook says hire before process. The AI-native playbook flips it.",
image_url:
"https://dealpulse-prod.hackathon.markster.io/assets/headshots/p3.png",
status: "draft",
},
},
],
requestId,
},
};
}
case "content_post_workspace":
case "content_post_workspace_get":
case "content_post_workspace_edit":
case "content_post_workspace_regenerate":
case "content_post_workspace_schedule": {
const modeFromSpec: string | undefined =
spec.name === "content_post_workspace_get"
? "get"
: spec.name === "content_post_workspace_edit"
? "edit"
: spec.name === "content_post_workspace_regenerate"
? "regenerate"
: spec.name === "content_post_workspace_schedule"
? "schedule"
: undefined;
const mode =
typeof args.mode === "string" ? args.mode : modeFromSpec ?? "get";
const postId = typeof args.post_id === "string" ? args.post_id : "9001";
if (mode === "get") {
return {
...base,
payload: {
workflow: "content_post_workspace",
mode: "get",
post_id: postId,
post: {
id: postId,
title:
"We grew Kontext Group from $71K to $1.3M. Here's what actually moved the needle.",
hook: "Everyone talks about AI replacing agencies. We used it to 18x a client's revenue.",
body: "In 2023, Kontext Group was doing $71K. Two years later, they crossed $1.3M. Not because we hired more people — we hired zero. Here's what we did differently...",
hashtags: ["#agencygrowth", "#ainative", "#revops"],
image_url:
"https://dealpulse-prod.hackathon.markster.io/assets/headshots/p2.png",
media: [
{
type: "image",
url: "https://dealpulse-prod.hackathon.markster.io/assets/headshots/p2.png",
},
],
status: "draft",
},
image_url:
"https://dealpulse-prod.hackathon.markster.io/assets/headshots/p2.png",
requestId,
},
};
}
if (mode === "edit") {
return {
...base,
payload: {
id: postId,
status: "draft",
body:
typeof args.body === "object" && args.body !== null
? (args.body as Record<string, unknown>).body ??
"Updated post body"
: "Updated post body",
updated_at: generatedAt,
requestId,
},
};
}
if (mode === "regenerate") {
return {
...base,
payload: {
id: postId,
status: "draft",
image_url:
"https://dealpulse-prod.hackathon.markster.io/assets/headshots/p3.png",
updated_at: generatedAt,
requestId,
},
};
}
if (mode === "approve" || mode === "schedule") {
return {
...base,
payload: {
id: postId,
status: "scheduled",
scheduled_for: "2026-02-21T09:30:00Z",
requestId,
},
};
}
return {
...base,
payload: {
item_id: sampleItemId,
generated_post: {
headline:
"Two founders. Zero employees. 41 customers. Here's the playbook.",
body: "We run Markster with Claude Code, Codex, and Cursor. No account managers. No SDR team. AI handles research, outreach, and content. We handle strategy and relationships.",
hashtags: ["#ainative", "#founderselling", "#scaleos"],
},
status: "scheduled",
scheduled_for: "2026-02-21T09:30:00Z",
requestId,
},
};
}
case "content_audio_library": {
return {
...base,
payload: {
tracks: [
{ id: "trk-1", name: "Studio Loop", duration_sec: 35 },
{ id: "trk-2", name: "Soft Rise", duration_sec: 42 },
],
requestId,
},
};
}
case "content_social_accounts": {
return {
...base,
payload: {
accounts: [
{
platform: "linkedin",
status: "connected",
handle: "@ivanivanka",
},
{ platform: "x", status: "connected", handle: "@TheIvanIvanka" },
{
platform: "email",
status: "connected",
handle: "ivan@markster.ai",
},
],
requestId,
},
};
}
case "tap_brand_assets_workspace":
case "outreach_brand_assets_workspace": {
return {
...base,
payload: {
brand: {
company: "Markster",
tagline: "4 agencies. 12 years. One system. Zero employees.",
positioning:
"AI-native growth agency. Full-stack: pipeline + outbound, content + thought leadership, proposals, CRM. One flat fee replaces your growth team.",
dinner_pitch:
"We built and exited 4 agencies over 12 years. Then we encoded everything we learned into a system. Zero employees - and our first client grew 9.2x in 15 months.",
offers: [
{
name: "Calibration Sprint",
price: "$5,000 one-time",
description: "14 days to learn your business and launch",
},
{
name: "Pipeline",
price: "$3,000-$5,000/mo",
description:
"Fill my calendar — 200-500 prospects/month, multi-channel outbound, meeting booking",
},
{
name: "Growth",
price: "$6,000-$9,000/mo",
description:
"Build my brand + fill my calendar — adds 8-12 LinkedIn posts/mo, 4-8 blog articles, content calendar",
},
{
name: "Scale",
price: "$10,000-$15,000/mo",
description:
"Be my growth department — ABM, event intel, CRM management, weekly strategy calls",
},
],
replacement_value:
"$225K/year (RevOps $100K + SDR $70K + Content Writer $35K + Tools $20K)",
},
social_proof: [
{
source: "First client",
result:
"9.2x revenue growth in 15 months. $71K → $1.3M. Proposals 2/week → 15+/week. Profit +31%.",
},
{
source: "BPO pilot",
result:
"Full pipeline built in 6 weeks. $105M revenue B2B industrial wholesaler, Burbank.",
},
{
source: "Organic demand",
result:
"477 companies on waitlist from running the same system on ourselves.",
},
],
testimonial: {
quote:
"Markster's AI does market research and proposals while I sleep. My team didn't grow - my system did.",
author: "Tamas Kiss, First Client",
},
differentiators: [
"Both founders code daily — Claude Code, Codex, Cursor",
"We run on the same system we deploy for clients (eating our own dog food)",
"No contracts — you stay because it keeps getting better",
"Compounding data: cross-client learning makes month 6 dramatically better than month 1",
],
requestId,
},
};
}
case "tap_icp_workspace":
case "outreach_icp_workspace": {
const stage = typeof args.stage === "string" ? args.stage : "review";
return {
...base,
payload: {
stage,
icps: [
{
id: "icp-1",
name: "B2B Service Agencies",
revenue_range: "$500K-$5M",
headcount_min: 5,
headcount_max: 50,
geo: "US (beachhead: Bay Area / SF)",
titles: [
"Agency Owner",
"CEO",
"VP Revenue",
"Head of Growth",
"Managing Director",
],
pain: "Founder IS the sales team. Spending 40-60 hrs/month on GTM. Feast-or-famine pipeline.",
prior_attempts:
"Tried agencies (black boxes, junior employees). Tried tools (still managing software).",
budget: "$3-10K/mo already being spent on marketing",
},
{
id: "icp-2",
name: "IT Services / MSPs (Beachhead)",
revenue_range: "$1M-$10M",
headcount_min: 10,
headcount_max: 50,
geo: "US (4,500+ in Bay Area, ~180,000 national)",
titles: ["Owner", "CEO", "VP Sales", "Marketing Director"],
pain: "Sell essential but boring services. Technical founders = AI-positive. Already budget $3-10K/mo.",
prior_attempts:
"Fear-based buyers (cybersecurity, compliance). Short sales cycles (1-3 months).",
budget: "$3-10K/mo",
},
],
disqualification: [
"< $300K revenue (can't sustain $3K+/month)",
"Solo founder with no clients yet",
"B2C business (our playbook is B2B)",
"Wants to buy a tool, not a service",
'"Can you do this for $500/month?"',
],
requestId,
},
};
}
case "tap_sequence_workspace":
case "outreach_sequence_workspace": {
return {
...base,
payload: {
sequence_id: sampleSequenceId,
name: "SF B2B Agency Owners — Peer Growth",
status: "active",
target_icp: "B2B Service Agencies ($500K-$5M)",
enrolled: 48,
steps: [
{
id: "step-1",
day: 0,
channel: "email",
subject: "Quick question about [company] growth",
status: "sent",
description:
"Warm intro — peer founder angle, specific observation about their business",
},
{
id: "step-2",
day: 2,
channel: "linkedin",
subject: "Connection request + note",
status: "sent",
description:
"Personalized LinkedIn connection with shared context",
},
{
id: "step-3",
day: 4,
channel: "email",
subject: "Thought this might help — [relevant case study]",
status: "sent",
description: "Value-first follow up with relevant proof point",
},
{
id: "step-4",
day: 7,
channel: "email",
subject: "15 min this week?",
status: "scheduled",
description: "Soft CTA — 'Would you be open to a brief call?'",
},
{
id: "step-5",
day: 14,
channel: "email",
subject: "One last thing",
status: "scheduled",
description:
"Final touch — share a relevant insight, no pressure",
},
],
analytics: {
sent: 180,
opened: 112,
open_rate: "62.2%",
clicked: 41,
click_rate: "22.8%",
replied: 16,
response_rate: "8.9%",
booked_calls: 5,
conversion_rate: "2.8%",
},
requestId,
},
};
}
case "tap_enrollment":
case "outreach_enrollment": {
return {
...base,
payload: {
sequence_id: sampleSequenceId,
enrolled: sampleLeadIds,
duplicate_rejected: Math.max(sampleLeadIds.length - 1, 0),
requestId,
},
};
}
case "tap_outreach_insights": {
return {
...base,
payload: {
events: [
{ event: "reply", count: 7, date: "2026-02-18" },
{ event: "email_open", count: 34, date: "2026-02-18" },
],
rollups: {
sent: 180,
opened: 112,
clicked: 41,
replied: 16,
booked_calls: 5,
},
requestId,
},
};
}
case "domain_mailbox_manage": {
return {
...base,
payload: {
domains: [
{
domain: "markster.ai",
dkim: "passed",
spf: "passed",
dmarc: "passed",
},
],
mailboxes: [
{
email: "ivan@markster.ai",
inbox_health: "healthy",
capacity_pct: 24,
},
],
requestId,
},
};
}
case "cold_mail_health_domains": {
return {
...base,
payload: [
{
id: 41,
customer_id: 13,
domain_name: "markster.ai",
status: "active",
health_score: 94,
bounce_rate: 1.2,
last_health_check: generatedAt,
created_at: "2025-11-01T10:00:00Z",
},
{
id: 42,
customer_id: 13,
domain_name: "go.markster.ai",
status: "active",
health_score: 91,
bounce_rate: 1.5,
last_health_check: generatedAt,
created_at: "2025-11-01T10:05:00Z",
},
],
};
}
case "cold_mail_health_mailboxes": {
return {
...base,
payload: [
{
id: 101,
customer_id: 13,
domain_id: 41,
email_address: "ivan@markster.ai",
status: "active",
health_score: 93,
bounce_rate: 1.1,
last_health_check: generatedAt,
created_at: "2025-11-01T10:10:00Z",
},
{
id: 102,
customer_id: 13,
domain_id: 42,
email_address: "growth@markster.ai",
status: "active",
health_score: 90,
bounce_rate: 1.4,
last_health_check: generatedAt,
created_at: "2025-11-01T10:12:00Z",
},
],
};
}
case "cold_mail_health": {
return {
...base,
payload: {
setup_status: "ready",
warmed_accounts: 5,
target_accounts: 5,
warmup_progress_pct: 100,
warmup_day: 21,
warmup_target_days: 21,
daily_send_capacity: 150,
blockers: [],
domain_health: {
domain: "markster.ai",
dkim: "passed",
spf: "passed",
dmarc: "passed",
blacklist_status: "clean",
sender_reputation: 94,
},
mailboxes: [
{
email: "ivan@markster.ai",
status: "healthy",
daily_capacity: 50,
sent_today: 12,
},
{
email: "hello@markster.ai",
status: "healthy",
daily_capacity: 50,
sent_today: 8,
},
{
email: "growth@markster.ai",
status: "healthy",
daily_capacity: 50,
sent_today: 15,
},
],
deliverability: {
inbox_rate_pct: 94.2,
spam_rate_pct: 3.1,
bounce_rate_pct: 2.7,
},
requestId,
},
};
}
case "customer_image_assets": {
return {
...base,
payload: {
uploaded: true,
assets: [
{
id: `asset-${randomSuffix(JSON.stringify(args))}`,
kind: "logo",
url: "https://cdn.markster.ai/mock-assets/logo.png",
},
],
requestId,
},
};
}
case "voice_pack_workspace": {
return {
...base,
payload: {
default_pack: "Founder Direct",
packs: [
{
id: "vp-01",
name: "Founder Direct",
tone: "specific, conversational, experience-backed",
description:
"Leads with facts and numbers, not questions. Uses 'I' freely. Drops into second person when teaching. Parenthetical asides for specificity. Writes the way he talks.",
rules: [
"Lead with a specific fact or number, never a question",
"Pattern: Statement. Evidence. Implication.",
"Average sentence length: 22 words (15-30 range)",
"Names, numbers, timeframes everywhere — never 'a client' or 'a company'",
"Zero banned words: leverage, utilize, innovative, synergy, game-changer, ecosystem",
"Never end with 'Agree?' or engagement bait",
"USD only, never EUR",
],
sample:
"I replaced 7 tools at a $4.4M company last month. Total monthly cost of their old stack: $4,133. Meetings booked in 90 days: 12. I cost them $997. I booked 23 meetings in 48 days.",
},
{
id: "vp-02",
name: "Community Vulnerable",
tone: "raw, generous, stream-of-consciousness",
description:
"Used in private groups and trusted spaces. Real pain, real numbers, real contrast. Never performative. Vulnerability earns credibility.",
sample:
"I exited from my last agency with 52 FTEs and 600+ contractors. I left because I couldn't keep up, burned out, and developed serious health problems. Today, the picture is extremely different: with my co-founder, we reduced our team headcount to 0.",
},
{
id: "vp-03",
name: "Twitter Sharp",
tone: "casual, sarcastic, technically generous",
description:
"Sharper, more casual. Detailed technical breakdowns. Dry observations. Sarcasm welcome.",
sample:
"Sven is moving too fast. He should really take another 38 months to make sure that 15k is safe, especially in such FX environment.",
},
],
signature_phrases: [
"Amateurs hustle. Pros systemize.",
"B2B growth without the corporate BS.",
"We build scalable predictable growth engines.",
"4 agencies. 12 years. One system. Zero employees.",
],
requestId,
},
};
}
case "pipeline_workspace": {
return {
...base,
payload: {
pipeline_summary: {
total_value: "$1,247,500",
weighted_value: "$487,250",
open_deals: 12,
avg_deal_size: "$103,958",
avg_cycle_days: 34,
win_rate_pct: 34,
},
stages: [
{ name: "Prospect", count: 4, value: "$285,000" },
{ name: "Discovery", count: 3, value: "$195,000" },
{ name: "Proposal", count: 2, value: "$167,500" },
{ name: "Negotiation", count: 2, value: "$250,000" },
{ name: "Closed Won (Feb)", count: 1, value: "$50,000" },
],
deals: [
{
id: "deal-1",
company: "Stripe",
contact: "Claire Hughes",
value: 50000,
stage: "Negotiation",
probability: 68,
days_in_stage: 5,
next_action: "Schedule pricing walkthrough",
last_activity: "2026-02-20",
},
{
id: "deal-2",
company: "Anthropic",
contact: "Sarah Chen",
value: 120000,
stage: "Proposal",
probability: 55,
days_in_stage: 8,
next_action: "Send enterprise proposal",
last_activity: "2026-02-19",
},
{
id: "deal-3",
company: "OpenAI",
contact: "Mina Patel",
value: 95000,
stage: "Discovery",
probability: 40,
days_in_stage: 12,
next_action: "Partnership alignment call",
last_activity: "2026-02-20",
},
{
id: "deal-4",
company: "Clay",
contact: "Maya Chen",
value: 67500,
stage: "Proposal",
probability: 60,
days_in_stage: 4,
next_action: "Follow up on ROI analysis",
last_activity: "2026-02-20",
},
{
id: "deal-5",
company: "Mercury",
contact: "Emily Hart",
value: 35000,
stage: "Discovery",
probability: 45,
days_in_stage: 7,
next_action: "Demo follow-up",
last_activity: "2026-02-18",
},
{
id: "deal-6",
company: "Ramp",
contact: "Daniel Brooks",
value: 40000,
stage: "Prospect",
probability: 25,
days_in_stage: 10,
next_action: "Initial outreach sequence",
last_activity: "2026-02-19",
},
{
id: "deal-7",
company: "Figma",
contact: "Reema Batta",
value: 150000,
stage: "Negotiation",
probability: 72,
days_in_stage: 6,
next_action: "Send pricing proposal — intent 87, ready to close",
last_activity: "2026-02-21",
},
{
id: "deal-8",
company: "Notion",
contact: "Ivan Ivanov",
value: 30000,
stage: "Prospect",
probability: 20,
days_in_stage: 6,
next_action: "Warm intro via shared connection",
last_activity: "2026-02-20",
},
{
id: "deal-9",
company: "Manufact",
contact: "Pietro Zullo",
value: 27500,
stage: "Closed Won",
probability: 100,
days_in_stage: 0,
next_action: "Onboarding kickoff",
last_activity: "2026-02-20",
closed_date: "2026-02-15",
},
{
id: "deal-10",
company: "Datadog",
contact: "Alex Kim",
value: 120000,
stage: "Discovery",
probability: 35,
days_in_stage: 9,
next_action: "Technical deep-dive call",
last_activity: "2026-02-19",
},
{
id: "deal-11",
company: "Scale AI",
contact: "Rachel Torres",
value: 95000,
stage: "Prospect",
probability: 20,
days_in_stage: 5,
next_action: "Initial outreach via warm intro",
last_activity: "2026-02-20",
},
{
id: "deal-12",
company: "Vercel",
contact: "Tom Nguyen",
value: 70000,
stage: "Prospect",
probability: 15,
days_in_stage: 3,
next_action: "Research + personalized approach",
last_activity: "2026-02-21",
},
],
at_risk: [],
requestId,
},
};
}
case "company_profile_workspace":
case "company_profile_workspace_get": {
const companyName =
typeof args.name === "string"
? args.name
: typeof args.company === "string"
? args.company
: null;
const knownCompanies: Record<string, Record<string, unknown>> = {
stripe: {
id: "comp-stripe",
name: "Stripe",
domain: "stripe.com",
industry: "Fintech / Payments",
city: "San Francisco",
employees: "5,200+",
valuation: "$95B",
founded: 2010,
fit_score: 72,
fit_reason:
"Large enterprise — above typical ICP range but strong expansion opportunity. Payment infrastructure partner potential.",
key_contacts: [
{
name: "Claire Hughes",
title: "VP Revenue",
intent: 89,
status: "Ready to Buy",
},
{
name: "Diana Lin",
title: "Head of Enterprise",
intent: 73,
status: "Hot",
},
],
},
anthropic: {
id: "comp-anthropic",
name: "Anthropic",
domain: "anthropic.com",
industry: "AI / Safety",
city: "San Francisco",
employees: "340+",
valuation: "$7B",
founded: 2021,
fit_score: 65,
fit_reason:
"AI infrastructure company. Partnership opportunity more than client fit.",
key_contacts: [
{
name: "Sarah Chen",
title: "VP Revenue",
intent: 88,
status: "Ready to Buy",
},
{ name: "Dario Amodei", title: "CEO", intent: 76, status: "Hot" },
],
},
openai: {
id: "comp-openai",
name: "OpenAI",
domain: "openai.com",
industry: "AI Infrastructure",
city: "San Francisco",
employees: "1,800+",
valuation: "$80B",
founded: 2015,
fit_score: 58,
fit_reason:
"Strategic partnership target. Platform distribution opportunity.",
key_contacts: [
{
name: "Mina Patel",
title: "Director of Partnerships",
intent: 82,
status: "Ready to Buy",
},
{ name: "Sam Altman", title: "CEO", intent: 57, status: "Warming" },
],
},
clay: {
id: "comp-clay",
name: "Clay",
domain: "clay.com",
industry: "Sales Intelligence",
city: "New York",
employees: "200+",
valuation: "$3.1B",
founded: 2017,
fit_score: 45,
fit_reason:
"Competitor in enrichment space. Potential acquisition or integration play.",
key_contacts: [
{
name: "Maya Chen",
title: "CRO",
intent: 87,
status: "Ready to Buy",
},
],
},
mercury: {
id: "comp-mercury",
name: "Mercury",
domain: "mercury.com",
industry: "Banking / Fintech",
city: "San Francisco",
employees: "400+",
valuation: "$3.2B",
founded: 2019,
fit_score: 70,
fit_reason:
"Startup banking — overlapping customer base. Strong integration opportunity.",
key_contacts: [
{
name: "Emily Hart",
title: "VP Sales",
intent: 84,
status: "Ready to Buy",
},
],
},
ramp: {
id: "comp-ramp",
name: "Ramp",
domain: "ramp.com",
industry: "Corporate Finance",
city: "New York",
employees: "800+",
valuation: "$8.1B",
founded: 2019,
fit_score: 68,
fit_reason:
"Finance automation overlaps with our deal impact analysis. Strong integration fit.",
key_contacts: [
{ name: "Daniel Brooks", title: "CRO", intent: 70, status: "Hot" },
],
},
figma: {
id: "comp-figma",
name: "Figma",
domain: "figma.com",
industry: "Design Tools",
city: "San Francisco",
employees: "1,200+",
valuation: "$12.5B",
founded: 2012,
fit_score: 55,
fit_reason:
"Large design platform. GTM team expansion indicates potential.",
key_contacts: [
{
name: "Sasha Ivanov",
title: "VP Revenue",
intent: 68,
status: "Hot",
},
],
},
notion: {
id: "comp-notion",
name: "Notion",
domain: "notion.so",
industry: "Productivity",
city: "San Francisco",
employees: "700+",
valuation: "$10B",
founded: 2013,
fit_score: 62,
fit_reason: "Workspace platform — strong complementary positioning.",
key_contacts: [
{ name: "Ivan Ivanov", title: "CRO", intent: 69, status: "Hot" },
],
},
manufact: {
id: "comp-manufact",
name: "Manufact",
domain: "manufact.com",
industry: "Developer Tools / MCP",
city: "San Francisco",
employees: "15",
valuation: "YC S25",
founded: 2024,
fit_score: 88,
fit_reason:
"Ideal ICP fit. Early-stage, founder-led, needs growth infrastructure. Active partnership.",
key_contacts: [
{
name: "Pietro Zullo",
title: "Founder",
intent: 74,
status: "Hot",
},
{
name: "Luigi Pignone",
title: "Co-Founder",
intent: 81,
status: "Ready to Buy",
},
],
},
};
const key = companyName?.toLowerCase().replace(/\s+/g, "") ?? "";
const match = knownCompanies[key];
if (match) {
return { ...base, payload: { ...match, requestId } };
}
return {
...base,
payload: {
id: `comp-${randomSuffix(key)}`,
name: companyName ?? "Unknown",
domain: "unknown.com",
industry: "Unknown",
fit_score: 0,
fit_reason:
"Company not found in CRM. Use sales_intelligence with mode=research to enrich.",
key_contacts: [],
requestId,
},
};
}
default: {
return {
...base,
payload: {
message: `Mock payload for ${spec.name} is not fully modeled yet.`,
requestId,
args,
},
};
}
}
}
