Skill Gate and Selection for an Agent Request

Request one, the gate and the wide ranking:

{
  "model": "jev-latest",
  "state": {
    "request": "{{user-request}}",
    "recent_context": "{{recent-context}}"
  },
  "questions": {
    "acts_on_things": {
      "type": "noul",
      "instructions": "Is the assistant being asked to act on the user's files, accounts, devices, or online services, rather than only to explain or advise?"
    },
    "needs_procedure": {
      "type": "noul",
      "instructions": "Would a careful expert answering this consult a specific documented procedure or set of commands, rather than answering from general understanding?"
    },
    "prose_suffices": {
      "type": "noul",
      "instructions": "Could a knowledgeable generalist fully satisfy this request in prose, with no tools, no documentation, and no access to the user's files or accounts?"
    },
    "skill": {
      "type": "choice",
      "instructions": "Which of these skills, if any, is the right one to load to help with the user's latest request?",
      "criteria": {
        "{{skill-1-name}}": "{{skill-1-description}}",
        "{{skill-2-name}}": "{{skill-2-description}}",
        "{{skill-3-name}}": "{{skill-3-description}}",
        "none": "No listed skill applies"
      }
    }
  }
}

Gate in code: average acts_on_things, needs_procedure and one minus prose_suffices; below 0.30 load nothing. Otherwise take the top three from skill.probabilities and send request two, one Noul per candidate, same state: "Does the skill '<name>' do the specific thing the user's request asks for? Read what it actually does, not just its name." Load the candidate with the highest answer only if it is above 0.30. Two small requests replace stuffing every skill description into the system prompt.