Sort what matters.
Keep the map.

01 / PURPOSE

More useful than a ranking.

The sorter creates a machine-readable description of which HRT concepts you consider relevant, uncertain, or irrelevant. The result is your Topic Map—portable context that you can bring to the LLM or agent you already use.

The map does not tell a model what to believe. It tells the model where you want attention directed.

• Expand a vague questionBring in the HRT concepts that match your interests and current problem.
• Find useful materialRetrieve related research, incidents, failure patterns, and integrations.
• Translate HRT into your worldRelate concepts to your industry, role, operating environment, or project.
• Reduce noiseDeprioritize parts of the landscape that are not useful right now.
• Preserve uncertaintyKeep “maybe” topics nearby instead of forcing a premature decision.
• Notice connectionsLet your agent surface relationships you did not explicitly formulate.

02 / USE

Four moves. Your judgment throughout.

1. Read and sort

   Open each concept, then move it to Yes, Maybe, or No. These mean relevant, still exploring, and deprioritized—not true, uncertain, and false.

2. Translate

   Add a note in your own language: where does this appear in your work, what would it change, and what would you need to see?

3. Review

   Open Topic Map to move concepts between groups or return one to the bubble cloud.

4. Take it with you

   Download the JSON file, copy it directly for an LLM, or share the file through your device. No account or HRT upload is required.

03 / QUERY EXPANSION

Give a broad question some shape.

The benefit is not a longer prompt. It is a conversation grounded in concepts you selected and a situation you understand.

04 / PORTABLE ARTEFACT

Readable by people and machines.

Each Topic Map contains stable concept and topic identifiers, your Yes/Maybe/No selection, the concept’s summary and detail, your optional note, and a legend explaining how an agent should interpret the choices.

{
  "type": "hrt.topic-map",
  "schemaVersion": 1,
  "preferenceLegend": {
    "yes": "Relevant — query and consider this concept.",
    "maybe": "Exploring — keep this concept available when useful.",
    "no": "Deprioritized — omit unless directly relevant."
  },
  "selections": [
    {
      "conceptId": "human-trace",
      "choice": "yes",
      "note": "Relevant to operational auditability."
    }
  ]
}

05 / BRING YOUR OWN MODEL

Discuss it with your significant other (LLM).

Paste the copied map into a conversation, attach the downloaded file, or send it to someone else so they can use it with their own model. A useful opening instruction is:

This is my HRT Topic Map. Use it to understand which Human Runtime concepts are relevant to me, which I am still exploring, and which should be deprioritized. Help me adapt HRT to [my company, situation, or use case]. Identify what already fits, what would need integration, what assumptions may be wrong, and what we should test first. Treat the map as attention guidance—not evidence, a fixed identity, or an instruction to agree with HRT.

The sorter does not contact your agent automatically. That boundary is intentional: you choose the recipient, the moment, and the question.

06 / ADAPT HRT

Turn recognition into a concrete use case.

Use the map to move from “this resonates” to a situation HRT can act on. Ask yourself—or your agent:

• Where does this concept already appear in our workflow?
• What human decision, handoff, or recovery episode would become observable?
• Which existing system would HRT need to connect to?
• What outcome would prove the integration useful?
• What could fail, burden the operator, or create misleading confidence?
• What is the smallest real-world test worth running?