synthlabs-ui-operator
$
npx mdskill add mkurman/zorai/synthlabs-ui-operatorNavigate SynthLabs visually for verifier review and settings.
- Enables verifier review, data preview, and settings changes.
- Depends on browser tools after synthlabs-setup confirms health.
- Executes browser-led workflows instead of backend automation.
- Delivers results through the visual app interface.
SKILL.md
.github/skills/synthlabs-ui-operatorView on GitHub ↗
--- name: synthlabs-ui-operator description: Use when SynthLabs work must be driven through the visual app for verifier review, data preview, settings changes, DEEP mode, or other browser-led workflows. tags: [synthlabs, synthlabs-ui-operator] --- # SynthLabs UI Operator ## Overview Use this skill when SynthLabs is best operated as an application instead of a backend route. Assume `synthlabs-setup` already confirmed a healthy instance before you begin. This skill does not provide dedicated zorai SynthLabs tools; it maps the work onto zorai's existing browser workflow. ## When to Use Use this skill when: - the task requires verifier review, manual approval, or row-level inspection, - the operator needs data preview or visible confirmation of dataset state, - settings changes or settings inspection must happen in the app, - DEEP mode is part of the requested workflow, - or the operator explicitly wants browser-led operation instead of backend automation. Do not use this skill when: - the task is session discovery, session creation, or repeatable generation that fits backend routes, - the task is autoscore, rewrite, remove-items, migrate-reasoning, or job polling, - or the local SynthLabs instance still needs to be started or repaired. ## Agent Rules - Use browser tools only after `synthlabs-setup` confirmed SynthLabs is healthy. - Prefer UI navigation for verifier review, data preview, settings changes, and DEEP mode. - Follow the browser workflow in `skills/operating/browser.md`: start by opening a browser pane, read the DOM before interacting, use element discovery instead of guessing selectors, prefer text-based clicks when possible, and read the DOM again after navigation. - Use existing browser tooling such as `open_canvas_browser`, `browser_read_dom`, `browser_get_elements`, `browser_click`, and `browser_type`; do not imply SynthLabs-specific zorai tools exist. - Do not use the browser just to confirm data or trigger work that stable backend routes already cover. - Hand session CRUD and repeatable generation work back to `synthlabs-generation`. - Hand autoscore, rewrite, remove-items, migrate-reasoning, orphan checks, and job polling back to `synthlabs-curation`. ## Browser-Led Workflow 1. Confirm the SynthLabs URL from the operator or from the healthy local instance already validated by `synthlabs-setup`. 2. Open the app in a browser pane and read the page before clicking anything. 3. Discover the relevant controls for the target surface: Verifier, Data Preview, Settings, Generator/Engine, or DEEP mode. 4. Capture the visible state you are reviewing or changing so follow-up work can resume from the correct session or screen. 5. If the task becomes repeatable through `/api/sessions`, `/api/ai/generate`, or `/api/jobs`, stop using the browser and switch to the backend-focused skill that owns that route. ## Owned Visual Surfaces ### Generator And Engine Screens - inspect visible session configuration, - review prompt fields and generation options before a run, - confirm which mode or screen the operator is using, - and capture any visual state that should be handed back to backend generation work later. ### Verifier Review - inspect rows, - compare outputs, - approve or reject entries, - and record issues when quality judgment depends on visual review. ### Data Preview - browse tables, - inspect row details, - confirm column meaning, - and visually validate imported or generated content. ### Settings Changes - update provider settings, - inspect DB mode or storage settings, - adjust visible workflow options, - and confirm that a settings change took effect in the app. ### DEEP Mode And Other Visual Workflows - navigate DEEP mode when the operator needs visible state and intermediate steps, - inspect mode-specific controls or panels, - and keep visually complex flows in the browser until they can be handed back to a backend-repeatable path. - capture enough visual state to hand the work back to `synthlabs-generation` or `synthlabs-curation` when the UI-only phase is complete. ## Handoff Rules - If the task can be expressed as `GET /api/sessions`, `POST /api/sessions`, or `POST /api/ai/generate`, switch to `synthlabs-generation`. - If the task can be expressed as `GET /api/jobs`, `GET /api/jobs/:id`, or a `/api/jobs/*` curation route, switch to `synthlabs-curation`. - Stay in this skill only while the UI is the real control surface. ## Common Mistakes - Opening the browser just to confirm data already exposed by `/api/sessions` or `/api/jobs`. - Repeating setup instructions here instead of requiring `synthlabs-setup` first. - Treating DEEP mode as a promised backend route when the task is explicitly visual. - Blurring the boundary between manual verifier review and repeatable backend generation or curation.
More from mkurman/zorai
- account-management>
- agile-scrum>
- albumentationsFast image augmentation library (Albumentations). 70+ transforms for classification, segmentation, object detection, keypoints, and pose estimation. Optimized OpenCV-based pipeline with unified API across all CV tasks. Supports images, masks, bounding boxes, and keypoints simultaneously. Note: classic Albumentations (MIT) is no longer maintained; successor AlbumentationsX uses AGPL-3.0. For torchvision-native augmentations, use torchvision.transforms.v2.
- aml-complianceAnti-Money Laundering (AML) and Know Your Customer (KYC) compliance workflow. Sanctions screening, PEP detection, transaction monitoring, suspicious activity reporting (SAR), and OFAC compliance.
- anki-connectThis skill is for interacting with Anki through AnkiConnect, and should be used whenever a user asks to interact with Anki, including to read or modify decks, notes, cards, models, media, or sync operations.
- approval-checkpoint-long-taskCanonical long-task pack for daemon-managed work with deliberate approval checkpoints, status summaries, rollback notes, and mobile-safe governance-aware updates.
- auditing-goal-artifactsUse when reviewing recent zorai goal run outputs, closure markers, ledgers, or evidence bundles to judge whether completion is credible or to identify remaining uncertainty.
- autogenAutoGen (Microsoft) — multi-agent conversation framework. Agent-to-agent chat, code generation & execution, tool use, group chat, and human-in-the-loop. Build collaborative AI systems with specialized agents.
- backtraderPython backtesting framework for trading strategies. Data feeds, brokers, analyzers, and live trading support. Strategy development with commission models, slippage, and signal-based execution.
- beautiful-mermaidRender Mermaid diagrams as SVG and PNG using the Beautiful Mermaid library. Use when the user asks to render a Mermaid diagram.