What MCP Assistants Can Do In QAVRA

This reference explains the QAVRA MCP capabilities in plain language. Use it to understand what ChatGPT or another connected assistant can read, draft, and save.

Quick Map

User RequestWhat The Assistant Can DoSaves Changes?
”List my QAVRA courses”Show courses you can administer.No
”Find the course with discovery in the title”Search your admin course list by partial title.No
”Show me the course structure”Load title, description, scope, groups, objectives, version context, and Knowledge Profile summary where available.No
”What is missing from this structure?”Load course context and QAVRA guidance so the assistant can reason about gaps, overlaps, depth, and scope.No
”Draft a better structure”Load course context and drafting guidance so the assistant can produce a draft in chat.No
”Search approved knowledge”Search available QAVRA Knowledge Profiles and source-backed content.No
”Update this title”Save approved course title or description changes.Yes
”Update this scope”Save approved intent, audience, outcome, domain, or boundary changes.Yes
”Save this final structure”Save approved objective groups and objectives.Yes

Read Capabilities

Read capabilities help the assistant understand what exists in QAVRA.

Course List

Use when the admin does not know the course ID.

Example:

List all of my courses in QAVRA.

The assistant receives a bounded list of courses the signed-in user can administer. If there are many results, it can search again with a title query.

Use when the admin knows part of the title.

Example:

Find my course with onboarding in the title.

The admin does not need to provide the full title. If multiple courses match, the assistant should ask the admin to choose one.

Course Structure

Use before brainstorming, drafting, or saving.

Example:

Show me the course structure for B2B Discovery Call.

QAVRA returns the saved course structure, course details, version context, current groups and objectives, and Knowledge Profile summary where available.

Use when the admin or learner-facing workflow needs approved source-backed context.

Example:

Search QAVRA for our refund policy.

The assistant should use available Knowledge Profiles and source-backed results instead of guessing from general knowledge.

Brainstorm And Draft Capabilities

Brainstorming and drafting do not call QAVRA’s backend course generator. QAVRA returns the current structure, course details, Knowledge Profile summary where available, and QAVRA structure guidance. The assistant then reasons in the conversation.

Structure Ideas

Use for prompts such as:

What is missing from this course structure?
What should we explore or deepen?
Which objectives are too broad, duplicated, or out of scope?

The result is advice, not a saved change.

Structure Draft

Use when the admin wants a proposed replacement or revision.

Example:

Draft a revised structure using the QAVRA rules and Knowledge Profile. Do not save it yet.

The assistant should show the proposed groups and objectives for review. It should not save the draft until the admin clearly approves the exact version.

Save Capabilities

Save capabilities change QAVRA. They should only be used after clear admin approval.

Update Course Details

Can save approved changes to:

  • Title.
  • Description.
  • Intent.
  • Audience.
  • Outcome.
  • Domain.
  • Boundaries.

Examples:

Rename this course to Sales Discovery Fundamentals. Do not change the structure.
Save this revised audience and outcome only.

Commit Course Structure

Can save approved changes to:

  • Objective groups.
  • Objective order.
  • Objective text.
  • Group titles.

Examples:

I approve this final structure. Save it to QAVRA.
Save the latest draft structure exactly as shown.

Approval Rules

The assistant should save only when the admin uses clear approval language.

Clear approval:

I approve this. Save it. Save this final structure to QAVRA. Update the title only.

Not clear approval:

Looks good. Maybe use this. Can you make it better?

If approval is unclear, the assistant should ask before saving.

Version And Publishing Boundary

Course structure reads include the latest saved major and minor version context available to QAVRA.

Saving through MCP is not the same as publishing. After a save, an admin may still need to:

  • Review the course in QAVRA.
  • Regenerate or review learner-facing content.
  • Check scenario practice.
  • Check assessment and evidence coverage.
  • Publish a major version or minor objective update.
  • Update offerings, LMS delivery, SCORM exports, or marketplace listings.

What MCP Cannot Do Today

The current QAVRA MCP connection should not be treated as a replacement for the full admin dashboard. It does not directly publish courses, export SCORM, launch offerings, manage learners, configure branding kits, or move products to marketplace.