Skip to content
Skills

Teach your coding agent to build with VAKYN.

Ten skills for Claude Code, Codex, Cursor and OpenCode. They find the decisions a chat model makes in your app, move them to VAKYN, and check the result on your own data before you switch.
The skills on GitHubMIT. Ask in plain words; your agent picks the skill.
claude plugin marketplace add vakyn-ai/skillsclaude plugin install vakyn@vakyn-ai

Then run /reload-plugins in an open session, or start a new one.

Start

Find it, build it, prove it.

No VAKYN server yet? Ask your agent:

Set up VAKYN on this machine.

It picks the model file for your hardware and walks you through install, start and API key.

  1. I

    Find

    vakyn-research finds where VAKYN fits in your app and ranks the spots.

  2. II

    Build

    vakyn-convert, vakyn-workflow, vakyn-cascade or vakyn-judge builds it. vakyn-batch runs it over what you already have.

  3. III

    Prove

    vakyn-check proves it on your data before you switch, and again after every change.

The skills

Ten skills, one plugin.

What each one does, when to reach for it, something to ask, and what you get back. Or call one directly, such as /vakyn:vakyn-convert.

  1. 01

    vakyn

    Start

    Checks whether VAKYN fits, picks a model file for your machine, sets up the server and makes a first call.

    Reach for it when
    You haven't run VAKYN yet, or the server won't start, or it answers with an error.
    You get back
    A running server, an API key kept in VAKYN_API_KEY, and a first answer.
    Ask
    Set up VAKYN on this machine.
  2. 02

    vakyn-research

    Find

    Finds the decisions a chat model or a person makes in your app or process, and ranks them.

    Reach for it when
    You want to know where VAKYN would help before you change anything.
    You get back
    vakyn-research.md: each decision with its file and line, ranked by gain against risk and effort. No code changed, no calls made.
    Ask
    Where could VAKYN help in this repo?
  3. 03

    vakyn-questions

    Build

    Picks noul, choice or score, writes the options and levels, and fixes a question that answers badly.

    Reach for it when
    You're writing a request, or an answer looks wrong or unsure.
    You get back
    The cause, and a rewritten request that passes the pre-send checks.
    Ask
    Why did it pick billing for this message?
  4. 04

    vakyn-convert

    Build

    Turns a chat-model prompt or a tool list into VAKYN calls, with Python or TypeScript code.

    Reach for it when
    A prompt in your app sorts, routes, flags or rates, or a chat model picks which tool to call.
    You get back
    Every line of the prompt sorted, the request as a .json file, and the code change. Your chat model keeps running side by side until you switch.
    Ask
    Make VAKYN pick the tool in agent.py.
  5. 05

    vakyn-workflow

    Build

    Designs a new decision workflow from a task in plain words: route, triage, pick a tool, check a claim, rate with a rubric, pick the best match, spot risk, fill a form.

    Reach for it when
    You're building something new and know the task, not the questions.
    You get back
    The questions, a state builder, working code from the closest pattern, and a path for unsure answers.
    Ask
    Build a router for our returns emails.
  6. 06

    vakyn-cascade

    Build

    A cheap chat model drafts, VAKYN checks the draft against its sources, and only failed drafts go to the expensive model.

    Reach for it when
    A pipeline writes answers from sources, such as help articles or contracts, and the expensive model does all of it.
    You get back
    The draft check, the routing code around your current call, a side-by-side log and a cost-saved summary.
    Ask
    Let a cheap model draft our support answers and send only the failed drafts to the big one.
  7. 07

    vakyn-judge

    Build

    Grades chat-model or agent outputs against your rubric, lines the grades up with your experts, and runs in CI.

    Reach for it when
    You grade replies, answers or agent runs by hand, or with a chat model as the judge.
    You get back
    One question per criterion, grades checked against your experts' labels, and a CI gate.
    Ask
    Grade our support bot's replies with this rubric and gate the CI on it.
  8. 08

    vakyn-batch

    Build

    Runs a CSV, Parquet or SQL table through VAKYN: a pilot first, then the full run with resume and progress.

    Reach for it when
    You have rows to label, tag or score, and questions that work on a few of them.
    You get back
    A script in your project, answer and confidence columns next to your data, and a review file for the unsure rows. Your source data is never overwritten.
    Ask
    Label every row in reviews.csv by topic and sentiment.
  9. 09

    vakyn-migrate

    Build

    Moves code that calls a hosted decision API to your own server, and points framework adapters at it.

    Reach for it when
    Your code already calls a hosted decision API and you want it on your own servers.
    You get back
    Every call site with its file and line, the config change, a side-by-side comparison and a way back.
    Ask
    Move our decision calls to our own VAKYN server.
  10. 10

    vakyn-check

    Prove

    Tests VAKYN against your chat model on your records, reads side-by-side logs, sets thresholds, checks calibration and drift, and reviews the code before go-live.

    Reach for it when
    Before you switch, and again after every change to the model, the questions or the code.
    You get back
    How often VAKYN agrees with your chat model, where to set the threshold for unsure answers, and a go-live review.
    Ask
    Here's a week of side-by-side logs. Can we switch?
Every skill

What the skills always do.

  • Checks every request

    Right shapes and limits, described options, an escape option such as other or none, one question per proposition.

  • Keeps exact work in code

    Arithmetic, dates, thresholds and rules stay in your code. VAKYN answers the judgment.

  • Gives unsure answers a path

    A person, a chat model or a review queue.

  • Runs side by side first

    A converted app keeps using your chat model and logs both answers until you switch.

  • Never invents answers

    If your server can't be reached, the agent says so and still writes the request and the code.

  • Keeps your key out of the chat

    It lives in the VAKYN_API_KEY environment variable, not in the conversation.

Generated code uses the official SDKs for the Jev-compatible API. They work with VAKYN once their base URL points at it: https://api.vakyn.com for VAKYN MAX, or your own server. On VAKYN MAX, put a key from the console in VAKYN_API_KEY and set VAKYN_BASE_URL to the same address; the skills read both.