productivity▌

Unix-style shell for navigating and querying web pages as a filesystem. \n \n Navigate URLs with cd , list page elements with ls , extract content with cat , and filter with grep —all operating on cached, locally-parsed page content \n Supports natural language intent inference; commands like \"show me the first 5 links\" or \"what forms are on this page?\" work alongside formal shell syntax \n Fully asynchronous design with instant prompt return; background tasks handle fetching, parsing, and e

Prepare marimo notebooks for scheduled batch execution with CLI parameters and optional experiment tracking. \n \n Converts UI-based parameters into Pydantic models that support both interactive forms and command-line argument parsing \n Enables dual-mode execution: iterate with the UI, then run batch jobs via CLI with arguments like --sample-size 4096 --learning-rate 0.005 \n Optionally integrates Weights and Biases for experiment logging and parameter tracking \n Includes EnvConfig for managin

Initialize and configure assistant-ui chat UI in Next.js projects or create new apps from templates. \n \n Supports two workflows: init for existing Next.js projects and create for new apps, with six starter templates covering AI SDK, cloud persistence, Clerk auth, LangGraph, and MCP integrations \n CLI flags enable non-interactive setup ( --yes ), template selection ( -t ), preset configuration ( --preset ), and component registry additions ( add ) \n Reference documentation covers advanced run

team-assemble

Frameworks for running productive one-on-one meetings drawn from seven product leaders. \n \n Emphasizes coaching over advising: shift to curiosity and empower reports to solve problems rather than providing immediate answers \n Recommends dedicated deep-dive conversations on past (life story), future (dreams), and present (career action plan) to build long-term development \n Includes monitoring for joy and recovery as performance indicators; treat lack of daily joy as a risk requiring interven

Structured guidance for building production AI agent systems through effective context management and multi-agent architectures. \n \n Covers foundational context engineering concepts including attention degradation patterns, context poisoning, and signal-to-noise optimization for language models \n Provides architectural patterns for multi-agent coordination (supervisor, peer-to-peer, hierarchical), memory system design, and filesystem-based context management \n Includes operational excellence

21 production-ready scripts for iOS app testing, building, and automation via accessibility-driven navigation. \n \n Semantic element interaction using accessibility trees instead of pixel coordinates, reducing token cost from 1,600–6,300 per screenshot to 10–50 per query \n Covers five domains: build and testing, navigation and gestures, accessibility auditing, permissions and notifications, and simulator lifecycle management \n All scripts support --json for machine-readable output and auto-de

Static security analysis for GitHub Actions workflows invoking AI coding agents. \n \n Detects nine attack vectors where attacker-controlled input reaches AI agents in CI/CD pipelines, including env var intermediaries, direct expression injection, CLI data fetches, dangerous sandbox configs, and wildcard user allowlists \n Scans .github/workflows/ for Claude Code Action, Gemini CLI, OpenAI Codex, and GitHub AI Inference steps; resolves one level of composite actions and reusable workflows \n Cap

Guide teams through learning-focused post-mortems and retrospectives using frameworks from 11 product leaders. \n \n Helps frame exercises as blameless learning opportunities rather than blame-focused reviews, with emphasis on psychological safety and honest sharing \n Covers core principles including pre-mortems with kill criteria, reframing failure as growth, and institutionalizing weekly impact and learnings reviews \n Guides users through structured questioning to surface systemic insights,

Git worktrees enable checking out multiple branches simultaneously in separate directories, all sharing the same repository. Create a worktree instead of stashing changes or cloning separately.