productivity▌

Build evaluation pipelines for LangSmith with LLM-as-Judge and custom code evaluators. \n \n Three core components: creating evaluators (LLM-as-Judge or custom code), defining run functions to capture agent outputs and trajectories, and running evaluations locally or auto-running via uploaded evaluators \n Supports both offline evaluators (comparing run outputs to dataset examples) and online evaluators (real-time quality checks on production runs) \n Requires LangSmith API key and project confi

Initialize Railway projects, add services, and configure deployments with automatic project detection. \n \n Handles three workflows: creating new projects with railway init , linking existing projects with railway link , and adding services to linked projects with railway add --service \n Automatically detects project state via railway status --json and walks up directory trees to find parent project links, avoiding unnecessary re-initialization \n Supports scaffolding for static sites (Vite, A

Linux environment setup for Flutter desktop development with OS detection and toolchain validation. \n \n Automatically detects Debian/Ubuntu or ChromeOS and halts with instructions for non-apt systems \n Installs core dependencies (curl, git, unzip, libglu1-mesa) and Linux desktop toolchain (clang, cmake, ninja-build, GTK libraries) \n Prompts for optional IDE setup (VS Code or Android Studio via snap) \n Runs flutter doctor -v and flutter devices to validate the environment, automatically inst

Fetch Twitter/X post content by URL without JavaScript or authentication. \n \n Retrieves full post metadata including author, timestamp, text content, images, and thread replies via Jina.ai API \n Supports both individual tweet fetching and batch operations across x.com and twitter.com URLs \n Includes bundled Python and Bash scripts for single and multiple tweet retrieval \n Requires a free Jina API key set as an environment variable to function \n

Debug and fix failing GitHub Actions checks in pull requests using gh CLI inspection and log analysis. \n \n Inspects PR checks via GitHub CLI, fetches GitHub Actions logs, and summarizes failure context with actionable log snippets \n Requires prior gh authentication with repo and workflow scopes; verifies status before proceeding \n Drafts a fix plan (using create-plan skill if available, otherwise inline) and implements only after explicit user approval \n Treats external CI providers (Buildk

Manage Apple Reminders from the terminal with list, date, and output filtering. \n \n Create, view, complete, and delete reminders with support for custom lists, due dates, and time specifications \n Filter reminders by today, tomorrow, this week, overdue, or specific dates in multiple formats (YYYY-MM-DD, ISO 8601, natural language) \n Output results as JSON, plain text (TSV), or quiet mode for scripting and automation \n macOS-only; requires remindctl CLI installed via Homebrew and Reminders a

Query CPU, memory, network, and disk usage metrics for Railway services. \n \n Supports nine metric types including CPU usage, memory, network traffic (RX/TX), and disk usage across deployments and instances \n Query metrics for a single service or all services in an environment using optional groupBy parameters (by deployment, instance, region, or service) \n Requires environmentId from railway status --json ; serviceId is optional to retrieve metrics across all services \n Time-based queries w

Framework-based assessment of product-market fit using signals from 46 product leaders. \n \n Apply the Sean Ellis \"very disappointed\" survey as a leading PMF indicator, targeting 40% threshold before long-term retention data is available \n Diagnose stage through retention curves, reference customer counts, and customer pull signals; distinguish between vanity metrics and genuine PMF evidence \n Recognize PMF across four levels (nascent to extreme) with segment-specific fit; understand that P

Local markdown search engine supporting keyword, semantic, and hypothetical-document queries across indexed collections. \n \n Three query types: lex (BM25 keyword matching), vec (semantic/natural language), and hyde (hypothetical document answers) with optional auto-expansion and intent-based disambiguation \n Accessible via MCP tools ( query , get , multi_get ), CLI commands, and HTTP API on port 8181 \n Supports collection filtering, phrase matching, term exclusion, and result reranking via r

Web search with AI-optimized results, relevance scoring, and advanced filtering via Tavily API. \n \n Supports three topic types (general, news, finance) and two search depths (basic, advanced) with optional AI-generated answer summaries \n Includes domain filtering to include or exclude specific sites, time-range filtering (day, week, month, year), and configurable result counts \n Returns structured results with relevance scores, publication dates, and optional raw page content; supports both