dimitri-vs
- 15 skills
- 0 followers
- 1 week ago last updated
- ▌ Mistral Ocr · dimitri-vs bundleConvert PDFs, scans, and document images to clean Markdown via the Mistral OCR API, with figures described inline by a vision model. Use whenever a PDF or scanned document needs text extraction — accuracy matters, so prefer this over pdftotext/pdfplumber/tesseract or ad-hoc approaches. Fast (~seconds for 100 pages, vs ~30s/page for local marker) and cheap (~$0.005/page). Requires MISTRAL_API_KEY env var.
- ▌ Pydantic AI · dimitri-vs bundleBuild AI agents with Pydantic AI framework. Use when implementing agents with Gemini/Google, structured outputs, tools, streaming, multi-agent delegation, or FastAPI integration.
- ▌ Railway CLI · dimitri-vsDeploy and manage apps on Railway's cloud platform. Use when deploying, checking logs, managing environments, or running commands against Railway services.
- ▌
- ▌ Codex Review · dimitri-vs bundleGet an independent review from Codex CLI (OpenAI, native Windows) as a second opinion. Use after implementing features, completing plans, finishing research/analysis, or anytime an unbiased third-party perspective would help - including non-code contexts like decisions, strategies, parenting situations, financial plans, or advice quality-checks. Runs 1-5 minutes in background (longer for complex reviews). Invoke when user asks for a sanity check, second opinion, or Codex review.
- ▌ Web Research · dimitri-vs bundlePerform web research using OpenAI APIs. Fast mode uses gpt-5.5 with web_search and medium reasoning for quick lookups (~1-2 min). Normal/deep modes use gpt-5.6-sol with web_search + code interpreter at high/xhigh reasoning effort for comprehensive multi-step research. Invoke when user needs current web information or thorough research on a topic.
- ▌ Skill Creator · dimitri-vsCreate new Claude Code skills with proper structure and best practices. Use when building a new skill, understanding skill architecture, or deciding between skills vs prompts vs CLAUDE.md.
- ▌ Routine Builder · dimitri-vsDesign, create, and maintain Claude Code Desktop routines (local scheduled tasks). Use when building a new routine, migrating a scheduled task from another platform, debugging a broken routine, or doing a monthly tune-up on existing routines. Covers the two-layer architecture (deterministic Python + editorial Claude), SKILL.md authoring, schedule configuration, and common pitfalls.
- ▌ Interface Design · dimitri-vs bundleInterface design guidance for utilitarian apps — dashboards, tools, admin panels, data-heavy UIs. Use when building new pages/views, adding major UI components, or reviewing frontend design consistency. Component-library-first approach.
- ▌ Clerk · dimitri-vsClerk authentication integration guide. Use when implementing user auth, setting up Clerk for a new project, configuring sign-in/sign-up flows, or handling auth-related security concerns. Covers agency workflow for client handover.
- ▌ Speak · dimitri-vs bundleRead the assistant's last message (or any text) aloud via OpenAI TTS, opening the audio in a media player with pause/scrub controls. Invoke when the user asks to speak, read aloud, say it, narrate, or "read that to me" — typically after a dense or information-heavy response. Rewrites the message into a listening-friendly script before synthesis. Supports phone delivery via ntfy for remote/cloud sessions.
- ▌ Humanize · dimitri-vsStrip Claude writing tells (em-dashes, "Not X, but Y" contrastives, colon-declarations, delve/realm/seamless/robust lexical cluster, wrap-up closers, synthetic empathy, collective modality) from text via the globally-installed `humanize` CLI. Invoke when the user asks to humanize, de-AI-ify, de-Claude, clean up LLM-sounding text, or make output sound less like a bot. Also use proactively after generating prose that will be shown verbatim to a human (LinkedIn posts, emails, proposals, essays) when the content should read as authored by a real person rather than an assistant.
- ▌ Skeleton Docs · dimitri-vsSkeleton UI framework reference. Use when working on Svelte frontend components, debugging Skeleton styling issues, customizing components beyond basics, or iterating on UI implementation. Fetch latest docs for best practices.
- ▌ Playwright · dimitri-vsBrowser automation via Playwright CLI and MCP. Use when interacting with web pages — navigating, clicking, filling forms, taking screenshots, scraping content. Defaults to CLI (`@playwright/cli`) for token efficiency. Falls back to MCP for short exploratory sessions or sandboxed environments.
- ▌ Sentry CLI · dimitri-vsQuery Sentry issues and events via sentry-cli. Use when investigating errors, triaging issues, or resolving them programmatically.