Index of every page on this site: llms.txt
All skills
Run it yourself

Ask every independent TypeSafe Jev question in one POST /v1/systemone (speculative fan-out). Use for parallel questions, smart-home or compound commands, one-call turn assessment, or when the user runs /jev-fanout. Do not fire one HTTP call per question.

Use this Skill: https://skilld.dev/gh/cobusgreyling/jev/jev-fanout

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

SKILL.md

≈67 tokens for metadata: the name and description. ≈531 when used: this file.

Description uses 3.3% of example budget

Before choosing a Skill, your Agent reads its name and description. All available Skills share that space.

  • A shorter description leaves more room for other Skills. This entry exceeds our 1% size suggestion.

In our Claude Code example, all Skill names and descriptions share 8,000 characters. This Skill uses ≈267 characters, or 3.3%.

The 1% threshold is a size suggestion. Longer descriptions can still fit.

Your model, settings, and other Skills decide how much text your Agent can read.

Example settings and source

The example uses a 200k-token context and default Claude Code settings. The count includes the name, description, separators, and when_to_use when present. Codex also counts local file paths.

Skit's source and limits: Codex 0.160.1, Claude Code 2.1.292.

Jev speculative fan-out

Ask every independent judgment over the same state in one POST /v1/systemone. Questions run in parallel and cannot see each other's answers. Code consumes the ones that apply.

Primitives and auth: skills/jev/SKILL.md. Official pattern: https://docs.typesafe.ai/patterns/fan-out

Do this

  1. List the judgments the workflow might need, including speculative ones (category, compound, room, device, availability, action per device type).
  2. Put each in questions with complete instructions. Write the speculative premise into the question (If this is a light command, …). Question IDs are not sent to the model.
  3. Mix choice, score, and noul in that one request. Pin jev-1.13.0 once thresholds are tuned; jev-latest moves.
  4. In code, branch on the answers that apply; leave the rest unused.
  5. Make a second HTTP call only when the next options or next state cannot be built yet.

Do not

  • One question per HTTP call
  • Hide five judgments in one Choice
  • Assume later questions saw earlier answers

Shape

{
  "state": "I've been trying to connect Stripe for 3 days. I'm losing sales. Help ASAP.",
  "model": "jev-latest",
  "questions": {
    "department": {
      "type": "choice",
      "instructions": "Which team should handle this message?",
      "criteria": {
        "billing": "Payments, invoicing, refunds",
        "technical": "Bugs, outages, integrations",
        "other": "Neither fits"
      }
    },
    "frustration": {
      "type": "score",
      "instructions": "How frustrated does the author appear?",
      "criteria": ["Calm, just stating facts", "Frustrated but civil", "Very angry, strong language"]
    },
    "is_urgent": {
      "type": "noul",
      "instructions": "Does the message convey urgency or time-sensitivity?"
    }
  }
}

This repo

  • examples/04_parallel_fanout.py — Choice + Score + Noul in one call
  • jev_lab/home.py and /home — ~13 speculative questions, then the dispatcher acts
  • packages/js/src/turn.ts — npx jev turn (route + input guard, one call)
  • docs/smart-home.md

Source: SKILL.md on GitHub

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

No alerts26d3 checks · Risk SAFE
  • Gen Agent Trust Hub26d

    This skill provides instructions for optimizing AI judgments by batching multiple independent questions into a single parallel request to the TypeSafe Jev service. It describes the JSON structure, primitives, and refers to external documentation and example scripts.

  • Socket26d

    No alerts

  • Snyk26d

    Risk: LOW · No issues

Signed by skilld at 77b65c6. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 2 weeks ago.

Activeupdated last month

README badge

README badge for cobusgreyling/jev/jev-fanout