Works with the AI tools you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIGitHub CopilotGitHub CopilotGemini CLIGemini CLI+20 more

    Jev AI-Judge

    1

    Built for developers and agent builders who ship AI features.

    $17

    Secure checkout via Stripe

    30-day refund guarantee

    Converts to your local currency at checkout

    0 installsSecurity scanned
    Jev AI-Judge

    Jev AI-Judge

    Example session with this skill installed

    Judge this product announcement before we publish it. Use the Jev AI-Judge workflow: decompose each claim into atomic questions, check them against the sources below, and return structured verdicts with calibrated confidence.

    Draft: "Our new analytics tier processes 2M events per second at half the cost of competitors, and is SOC 2 certified."

    Sources: [1] Benchmark report: peak 2.1M events/s sustained over 24h. [2] SOC 2 Type II certificate, valid through 2027-03. (No pricing source attached.)

    • Read your context and instructions
    • Compiled the jev ai-judge

    Jev AI-Judge Verdict

    Overall: NEEDS REVISION — 2 of 3 claims supported, 1 unsupported, 0 safety flags.

    Claim verdicts

    #Atomic claimVerdictConfidenceEvidence
    1Processes 2M events per secondSUPPORTED0.92Source [1]: "peak 2.1M events/s sustained over 24h"
    2Half the cost of competitorsUNSUPPORTED0.88No pricing source attached; competitor pricing varies by workload — cannot verify
    3SOC 2 certifiedSUPPORTED0.81Source [2]: SOC 2 Type II certificate, valid through 2027-03

    Confidence tiers

    • ≥ 0.85 → auto-approve: claims 1
    • 0.60–0.85 → flag for human review: claim 3 (confirm the cert covers this product tier)
    • < 0.60 → block and escalate: none

    Recommended action

    Revise the draft before publishing: drop or reword the "half the cost of competitors" claim, or attach a pricing source and re-run the judge. All other claims are safe to ship.

    Example output. Judged in ~210ms at roughly 1/100th the cost of a frontier-LLM judge — cheap enough to verify every output instead of sampling.

    Connects securely to your tools. The creator never sees your data.

    What you get

    The 6-step Jev judging workflow: define the verdict, decompose into atomic questions, build an evidence pack, batch in one parallel call, branch on confidence tiers, log verdictsFour ready-made judge templates: hallucination/claim-support detection, safety triage, quality scoring, prompt-injection screeningThe agent gate pattern (generate → judge → revise or escalate), plus the anti-patterns that make judges fail

    About this skill

    Stop trusting your AI agent's output blindly. Jev AI-Judge teaches your agent to use Jev — TypeSafe's new System One model — as a dedicated verifier: state plus typed questions in, structured verdicts with calibrated confidence out, in ~70-500ms and at roughly 1/100th the cost of a frontier-LLM judge, so you can verify EVERY output instead of sampling. Inside: a 6-step judging workflow (define the verdict, decompose into atomic Choice/Score/Noul questions, build the state as an evidence pack, batch in one parallel call, branch on confidence tiers, log verdicts), four ready-made judge templates (hallucination/claim-support detection, safety triage, quality scoring, prompt-injection screening), the agent gate pattern (generate → judge → revise or escalate), and the anti-patterns that make judges fail. Built for developers and agent builders who ship AI features. Requirement: a TypeSafe API key (Jev is in early access at typesafe.ai) — the skill teaches the judging method and points to docs.typesafe.ai for the current API wire format.

    How to install

    Works the same in every agent - Claude, Cursor, Codex, Copilot and 20+ more.

    ~30 seconds
    1. 1

      Download the ZIP

      Free skills download straight away. Paid skills unlock right after purchase.

    2. 2

      Unzip into your skills folder

      Every agent reads skills from one folder on your machine. Drop the unzipped folder in there.

    3. 3

      Ask your agent to use it

      Restart the agent if it was already running. It picks the skill up automatically - no config needed.

    Skills folder by agent

    Click the path to copy it. Create the folder if it does not exist yet.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill passes our trust checks below.

    Security scanned

    Passed our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    30-day refund

    Not a fit? Get your money back

    Trust & safety

    Security scanned

    Verified clean 8 days ago

    • Passed all security checks, Safe to install

    Listed8 days ago

    What's inside

    Frequently Asked Questions