Skip to main content

Orchestrating Intelligence: Collaborating with Masters

In Masterminds, you are not prompting a chatbot; you are commanding a team of elite Silicon Valley esperts. We call them Masters.

Unlike standard AI that waits for your perfect instruction, Masters are architected to take initiative. They know the frameworks (Jobs-to-be-Done, Outcome-Driven Innovation, Lean Canvas) better than most human product managers. Your role shifts from "Prompt Engineer" to Product Director.

1. The Interaction Model: Propose, Validate, Refine​

The core loop of Masterminds is built on Proactive Intelligence.

The "Propose-First" Mandate​

When you engage a Master like Tony (Strategy) or Teresa (Solution), they will not just answer your questions. They will guide you through best-in-class product development frameworks, proposing strategic artifact based on frameworks' deliverables and sophisticated context engineering. Check this simplified examples on how master operates and how to interact with them:

  • Example: You say "I want to build a crypto app."
  • Standard AI: "Here are some features for a crypto app..."
  • Master Tony: "To de-risk this, we need to define the 'Underserved Outcome' first. Based on current market signals, let's draft an Outcome-Driven Roadmap that focuses on 'Trust' as the primary metric. Let's execute the process to produce this deliverable with the highest confidence."

Your Role: Validation & Challenge​

The Master do the heavy lifting for you in every step of the frameworks' process: research, analysis, insights, recommendations, you name it. They provide deliverables in each step, attached as a "chip" in the master's response message, that generally includes a TL;DR and call-to-actions for you. Now it your time to take action to validate their substanceor ask for refinements!

  • Challenge Assumptions: "Tony, you assumed our user is a 'Day Trader'. What if we pivot to 'Long-term Holder'? How does the roadmap change?"
  • Demand Evidence: "Teresa, this solution seems generic. Use MCP Web Search to find 3 competitors who tried this and failed."
  • Request Resources: "Eric, I'm not ready to approve this PRD. Run a Sentiment Analysis on these uploaded user interviews first."
  • Change Format: "Steve, provide me with a 3D visualization of this table."

Refining the Output​

Once the Master defends their proposal or incorporates your feedback, they update the Deliverable(s).

  • The Review: Click the deliverable chip to open it in the Visualizer.
  • Ask the Master to Revise Precisely: If you want the Master to adjust something, ask for a precise revision—be specific about what to change and reinforce, "Please update only section X, and keep everything else unchanged." This targeted approach helps prevent unintended edits elsewhere.
  • Manual Edit: If you want to change something manually—sometimes it's faster to do it yourself—just use the Editor. Your edits are committed directly to the chat's context when you save them.

2. Empowered Messaging: Steering the Ship​

Because Masters are proactive, they sometimes race ahead. The Communication Hub gives you the tools to steer them back on course without breaking the flow.

"Time Travel" (Regenerate From Here)​

If a Master goes down a strategic rabbit hole you disagree with, don't argue for 20 messages.

  • Action: Scroll up to the point of divergence (e.g., where Tony suggested "Subscription Model").
  • Command: Click Regenerate From Here.
  • Result: You prune that entire future. You can now steer Tony: "Actually, let's explore a Transaction-Fee model." The Master forgets the mistake and builds on the new direction instantly.

"Correcting the Record" (Edit)​

Masters treat your input as the definitive Ground Truth, so editing your prior message can significantly influence their reasoning. There are two ways to use Edit Message:

  1. Edit Only (No Regenerate):
    If you simply edit your previous message, the updated information will be visible to the Master in all future responses and deliverables, but the existing conversation history and artifacts remain unchanged. Use this when you want to add or clarify information for the next steps, without altering what’s already happened.

  2. Edit and Regenerate From Here:
    After editing your message, you may also click Regenerate From Here. This prompts the Master to re-run its entire reasoning chain from the corrected message onward, updating all dependent messages and deliverables. Use this if the error or change should affect the entire flow of conversation and outputs following the edit.

In both cases, precise corrections at the source help Masters deliver the highest quality product work.

3. Grounding the Masters (Fact-Gating)​

Masters are powerful, but they are not psychic. To get SV-grade output, you must feed them Ground Truth.

  • Upload Context: For example, don't just describe your users; upload the raw interview transcripts, or persona documents you may have.
  • Data Dependencies: A Master like Linus (Delivery) cannot build a robust system without the context from Teresa (Solution) and Jony (Product Design). Use Chat Context Import to ensure the "Chain of Thought" flows from Strategy to Execution.

4. Minds vs. Masters​

While Masters (like Eric, Tony or Linus) guide you with SV-Grade Processes, Minds (like Gump and Melissa) operate differently.

  • Minds are Creative Engines: They do not follow a rigid de-risking protocol.
  • Your Role: You must direct Minds more explicitly. "Gump, take Tony's roadmap and visualize it as a Sankey Diagram." "Melissa, take Jony's product design and recommend three subscription pricing tiers that match it."
  • Symbiosis: Masters set the Truth; Minds like Gump and Melissa build, visualize, or strategize creatively on top.

5. Instant Minds​

While executing a Master process or collaborating with a Mind, you sometimes need immediate perspective from other domain experts. Instant Minds let you summon targeted expertise—by role or specialty—right inside any conversation or artifact, without leaving your current context.

Anytime, you can @-invoke experts by role, such as @McKinseyStrategist or @Legal, or any relevant expert persona like @MartyCagan—without leaving your chat's context. Just tag the specialist you need and ask your question:

  • Example 1 (inside a Jony workflow):
    "@Legal: Are there any legal consequences to designing this product in this way?"

  • Example 2 (while considering strategy in a Master workflow):
    "@McKinseyStrategist: Will this pricing model impact my future revenue in any way?"

Instant Minds respond contextually, drawing on the current artifact, message, or workflow.

Note: currently you can't summon the Minds team (like Gump or Melissa) inside other chats.

When to Use:

  • To get legal, strategic, or compliance advice on a deliverable, in the moment.
  • To verify assumptions, spot risks, or get specialist recommendations—without derailing your work with process overhead.
  • To add multiperspective checks mid-flow, ensuring your work meets external or regulatory standards.

By instantly embedding external expert views right within your SV-grade process, you level up every decision—without breaking pace or context.

By understanding this dynamic—Masters proposing structure, you validating substance, and Minds illustrating the vision—you unlock the true power of the Integrated Product Environment.

6. Useful Commands​

Leverage these commands and controls to guide your workflow:

  • 📚 Help: Type help at any time (in any chat) to view product-development concepts or get tool assistance.
    Examples:

    help
    help: please explain what an OST stands for?
    help: what are key steps in Product Discovery?

  • ⚙️ Preferences: Set your language, verbosity, tone, or humor style for a more personalized experience.
    Examples:

    set language to Portuguese
    set verbosity to concise
    set tone to formal
    set humor on

  • 📣 Feedback: Click the Feedback button in the top left command bar to report bugs or request features.
    Examples (Feedback Prompt):

    "When trying to generate a roadmap, the suggestions panel doesn't load."
    "Feature request: Can I export a Mind's diagram to SVG?"

7. Vibe Mode (Alpha – Use With Care)​

Vibe Mode lets you fast-track or simulate steps in a Master workflow by letting agents fill in any missing pieces on your behalf.

How to use:
Type Vibe Mode ON in chat to activate. To limit scope, you can specify how many steps:
Turn Vibe Mode on for the next 2 steps.

What happens:

  • When Vibe Mode is enabled, agents will optimistically fill in the blanks for each Master action, using all available context to choose the most reasonable option for each step.
  • Agents do not "read your mind" or anticipate ambiguous intentions—they will only fill gaps where context makes your intent clear.
  • This mode is ideal for quickly moving through routine, boilerplate, or process-heavy steps, or for simulating possible workflow outcomes.

Caution:

  • Vibe Mode is alpha: Results may not match your exact expectations or catch subtle project nuances.
  • Best Practice: Only use Vibe Mode for specific steps—not for full workflows without review.
  • Always double-check outputs before relying on them for important or irreversible decisions. Turn off Vibe Mode when precision and hands-on control are required.

8. God Mode (Advanced – Expert Use Only)​

God Mode is an advanced operation mode that lets you override standard Master Agents' workflows, giving you direct, unstructured control over agent capabilities. Unlike Vibe Mode—which auto-fills routine workflow steps—God Mode lets you actively skip, re-order, or re-format steps, or even use agents as pure subject experts for actions they usually execute under its methodological framework.

How to Use​

  • Activate: Type God Mode ON in chat. You can specify what you want to override, skip, or direct, e.g.: 
    God Mode ON. Skip to step 4, using my provided personas.
    God Mode ON. Run only the Outcome Prioritization logic on this data.
    God Mode ON. Ignore the roadmap steps. Just do an Opportunity Solution Tree using this research.
    God Mode ON. Let me manually choose which steps (and inputs) to run, in which order.

What Happens​

  • With God Mode enabled, agents will no longer strictly follow the Masterminds’ framework or procedural step-by-step execution.
  • You can jump ahead, skip unnecessary steps (when you have already provided required context), or cherry-pick specific functions that the agent excels at and apply them independently.
  • Agents may accept flexible, compressed, or alternative input formats.
  • If you know the agent’s expertise (e.g., Opportunity Solution Trees, Outcome Roadmaps, Prioritization), you can invoke those capabilities directly—even out of order, or detached from the broader workflow.

When to Use (and NOT Use)​

God Mode is for advanced users who:

  • Have deep knowledge of the problem space or process, and want to experiment or move faster.
  • Wish to test downstream outcomes before completing prerequisite steps (provided context is sufficient).
  • Believe that changing the step order or skipping parts of a workflow will add value for their situation.
  • Want to leverage specific agent superpowers (e.g., "Apply only the Prioritization algorithm to this feature list.").

Example scenarios:

  • You’ve mapped an Outcome Roadmap outside Masterminds and wish to jump straight to Prioritization.
  • You want the agent to generate only the Opportunity Solution Tree based on research you already have—skipping the customer discovery steps.
  • You wish to remix or extend the planned workflow in non-linear ways.

⚠️ Caveats and Warnings​

  • Reduced Statistical Confidence: By leaving the framework designed for Masterminds agents, you’re also bypassing checks and balances that ensure statistical and experiential integrity. Answers given may no longer carry the same level of reliability, especially if critical prerequisites are skipped intentionally.
  • Loss of Guidance: You’re responsible for providing enough context and for checking the agent’s outputs. Gaps in logic or incomplete hand-offs are possible.
  • Best Practice: Use God Mode only when you are confident in the process and comfortable reviewing outputs critically. For decision-critical or complex flows, stick with standard or Vibe Modes.

Masterminds empowers you to bend the tools as you see fit, but always “trust, then verify.” When in doubt, leverage the standard flow or ask an agent for recommended usage.