Examples

Real launch packs from chiefmo_launch_product

Same JSON envelope your agent will receive. Copy the shape, build against it, ship. Each example is a real production response — generated by calling the live MCP endpoint.

The shape (annotated)

{
  "launchId": "<run-id>",                // pass to chiefmo_measure_launch_results 24h later
  "intent": "launch_product",
  "run": { ... },                            // full run record (status, cost, etc.)

  "launchPack": {
    "positioning": { assetId, body },        // your LLM renders body into final positioning
    "launchAngle": { assetId, body },        // launch narrative / angle
    "channels": {
      "linkedin":     { assetId, body, publishViaConnector: "zernio",  ... },
      "x":            { assetId, body, publishViaConnector: "zernio",  ... },
      "product_hunt": { assetId, body, publishViaConnector: "zernio",  ... },
      "email":        { assetId, body, publishViaConnector: "resend",  ... },
      "landing_hero": { assetId, body, publishViaConnector: "site_builder", ... }
    }
  },

  "generatedImages": [                       // base64 image data + cost
    { id, usedFor, status, generator, estimatedCostUsd, retailCredits }
  ],

  "publishActions": [                        // approval-gated; one per channel
    {
      "id": "launch-action-<runId>-linkedin",
      "channel": "linkedin",
      "connector": "zernio",
      "executorTool": "chiefmo_publish_approved_post",   // ← call THIS after approval
      "status": "draft",
      "preflight": { severity, warnings, recommendations, gates }
    },
    {
      "id": "launch-action-<runId>-email",
      "connector": "resend",
      "executorTool": "chiefmo_send_approved_email",     // ← email path
      ...
    }
  ],

  "reviewUrl": "https://chieflab.io/runs/<id>?token=...",  // HMAC, 7-day TTL, no login
  "trackingPlan": {
    "metricsTrackedAfter": "24h",
    "nextRunTool": "chiefmo_measure_launch_results"
  }
}

How agents typically consume this

// 1. Call launch
const r = await mcp.call("chiefmo_launch_product", { productUrl, goal });

// 2. Surface reviewUrl to the user (they approve in browser)
console.log(`Review here: ${r.reviewUrl}`);

// 3. Render each channel brief into final copy with YOUR LLM
for (const channel of Object.keys(r.launchPack.channels)) {
  const brief = r.launchPack.channels[channel].body;
  const finalCopy = await yourLLM(`Render this brief: ${brief}`);
  // Show finalCopy to the user, get final approval, store for step 4
}

// 4. After approval, fire each publishAction via its executorTool
for (const action of r.publishActions) {
  if (action.connector === "zernio") {
    await mcp.call("chiefmo_publish_approved_post", { actionId: action.id, content, platforms });
  } else if (action.connector === "resend") {
    await mcp.call("chiefmo_send_approved_email", { actionId: action.id, from, to, subject, html });
  }
}

// 5. Wait 24h, then:
const review = await mcp.call("chiefmo_measure_launch_results", { runId: r.launchId });
// review.brief contains the next-iteration recommendation

Want more examples? Email us with the product type / industry you're modeling.