A.M.A.N.D.A. // Supervised Decision Runtime

Six governed stages turn evidence into a decision a human can own.

A.M.A.N.D.A. — Adaptive Multi-Agent Neural Duality Analyzer — is Hyde Workshop's supervised six-stage decision system. It defines the mission, verifies evidence, audits the implementation, routes the work, tests the defenses, and prepares a recommendation that stops at a human approval gate.

  • Six supervised stages
  • Human approval required
  • Autonomous deployment disabled
  • Public page does not start a run
AI Astra representing the A.M.A.N.D.A. accountability and human approval stage
AI Astra // Accountability stage // The system may recommend; final consequential approval remains human-controlled.

Operational Proof // A.M.A.N.D.A. v1 Baseline

The framework now has a working governed runtime behind it.

In controlled administrator testing, the full six-stage chain reached HUMAN_REVIEW, waited for an explicit human decision, and completed only after approval. The runtime did not authorize autonomous deployment.

Six-stage handoff Amy, Maya, Aria, Nancy, Dawn, and Astra completed the governed chain.
Human review gate The workflow stopped at HUMAN_REVIEW instead of approving itself.
Human decision recorded An explicit human approval moved the governed run to COMPLETE.
Deployment still blocked Autonomous deployment remains unauthorized by design.
  1. AmyMission
  2. MayaEvidence
  3. AriaAudit
  4. NancyRouting
  5. DawnDefense
  6. AstraDecision
  7. HumanApprove / Revise / Reject

Hyde Workshop Difference // Supervised Decision Layer

Do not automate everything. Make the consequential parts accountable.

Hyde Workshop's position is not “AI replaces judgment.” The system separates Cain pressure from Abel value, preserves evidence and control boundaries, and identifies the human who owns the final action.

Cain // Pressure

What can fail?

Weak evidence, excessive permissions, hidden actions, privacy exposure, irreversible change, poor recovery, and false certainty.

Abel // Value

What creates measurable benefit?

Clearer decisions, useful output, accessible delivery, time saved, repeatable workflows, maintainable systems, and real customer value.

Human // Accountability

Who owns the final action?

Sources, assumptions, limits, approval gates, rollback, corrections, and the named decision owner remain visible.

Interactive Pipeline // Six Governed Stages

Each stage owns one responsibility. No stage owns the final human decision.

Select a stage to see its role, primary output, and control boundary. This interaction is local to the page and does not start an A.M.A.N.D.A. run.

AI Amy representing the A.M.A.N.D.A. Mission stage
Mission

AI Amy

Turns an unclear request into a defined mission with audience, objective, scope, constraints, acceptance criteria, and a named human owner.

Primary output: mission brief and acceptance criteria.

Planning Tool // Browser-Only Brief Builder

Describe the outcome. Generate a structured route—not a black-box promise.

This planning tool runs entirely in your browser. It does not call Gemini, Cloud Run, or the live A.M.A.N.D.A. runtime.

Use the generated brief as a starting instruction. Review it before sending it to Hyde Workshop or using it in a governed workflow.

WordPress / Elementor controlled repair

Define the defect, gather evidence, audit the smallest relevant layer, route the implementation, verify defenses and rollback, then stop for human approval.

Amy // Mission
Maya // Evidence
Aria // Audit
Nancy // Routing
Dawn // Defense
Astra // Decision

Example route loaded. Generate a route to customize the planning brief.

Where A.M.A.N.D.A. Fits

One governed method. Multiple practical workflows.

AI tool adoption

Compare workflow fit, evidence, cost, permissions, failure modes, and the human-control boundary before adoption.

WordPress / SEO

Separate page code from Elementor, WordPress, plugins, hosting, caching, analytics, and consent before changing production.

Security / privacy

Surface permissions, sensitive data, consent, unsafe actions, recovery paths, and rollback requirements before remediation.

Creative production

Define the brief, verify source material, review quality and provenance, protect rights, and keep final publication human-owned.

Compare Systems in AI Lab Read Current Signals Open AI Forge Try the A.M.A.N.D.A. Workshop

Control Standard // What A.M.A.N.D.A. Does Not Hide

Trust requires visible limits and recovery paths.

No automatic public run

Viewing this page, using the brief builder, or submitting the public contact form does not automatically start a billable A.M.A.N.D.A. run.

No autonomous consequential action

Publishing, deployment, spending, deletion, access changes, customer actions, and security decisions remain behind human approval.

No invisible certainty

Evidence gaps, blocked stages, revision requirements, assumptions, safeguards, and rollback remain part of the decision record.

Next Route // Governed Action

Compare first. Approve consequential action last.

Use AI Lab to compare systems, AI Forge to understand implementation controls, the A.M.A.N.D.A. Workshop to practice the mission-to-human-review sequence, and Contact when you need direct help turning a real problem into a governed implementation.