agent-foundry

This skill should be used when the user asks to "compress an image", "optimize images", "reduce image file size", "make images smaller", "convert to WebP", "convert to AVIF", "optimize images for the web", "create a srcset", "responsive images", "compress a GIF", "shrink a GIF", "batch compress a folder", "losslessly compress images", or wants to compare compression levels and preview before/after results. Uses caesiumclt, rimage, and gifsicle. Triggers include: "compress", "optimize image", "WebP", "AVIF", "srcset", "lossless", "reduce file size", "image too big", "gifsicle".

This skill should be used when the user asks to "generate an image", "create an image", "make a picture", "draw me a", "create artwork", "make an illustration", "edit an image", "modify a photo", "upload a file to Gemini", "check image stats", "change the aspect ratio", or mentions nanobanana, Gemini image generation, or AI art creation. Triggers include: "generate an image", "edit image", "create a picture", "4K image", "model tier", "aspect ratio", "negative prompt", "style transfer".

This skill should be used when the user asks to "harden npm config", "create .npmrc", "secure my npm setup", "configure lockfile-lint", "add preinstall script", "harden package.json", "generate security config", "add CI security checks", or "harden my project against supply chain attacks". Generates or updates configuration files only — does not install tools.

Use this skill to generate a Midnight Network DApp spec brief through an adaptive Q&A. The output is a concise brief (not a full spec) intended as input to a downstream spec / planning skill such as /superpowers:brainstorming or /sdd:specify. Triggers on "generate midnight dapp spec", "improve midnight dapp prompt", "midnight dapp brief", "midnight dapp prompt", or /prompt-rewriter:midnight-dapp-spec.

Plan steps that already know which command to invoke should call the command directly and skip this skill. Use this skill only when you need to pick a decision-making technique for the current question — it matches the decision shape against seven techniques (tribunal, adversarial, red-team, pre-mortem, council, ladder-of-abstraction, calibration) and recommends one.

This skill should be used when the user asks to "share a transcript", "export a session", "convert a transcript", "share this conversation", "create a shareable transcript", "export to HTML", "export to markdown", or mentions sharing Claude Code sessions with others.

This skill should be used when the user asks to "share a file", "upload to agentbin", "publish an HTML page", "host a rendered page", "get a public URL for a file", "share output publicly", or when an agent needs to make an HTML, Markdown, YAML, JSON, TOML, or XML file accessible at a public URL for linking in GitHub PR comments, issues, or similar.

This skill should be used when the user asks to "set up supply chain protection", "install security tools", "configure Socket", "install lockfile-lint", "set up supply chain defence", "fix supply chain issues", or when invoked by the doctor command to remediate missing tools. Handles installation and configuration of Socket.dev CLI, lockfile-lint, LavaMoat (optional), CycloneDX (optional), and project state file initialisation.

This skill should be used when the user asks to "audit dependencies", "run security audit", "check for vulnerabilities", "supply chain report", "security posture", "audit npm packages", "check dependency security", or "generate SBOM". Runs a comprehensive security posture report using npm audit, lockfile-lint, Socket.dev, and dependency age analysis. Cross-references devs:deps-core for package manager commands.

Create distinctive frontend interfaces with faux-ASCII aesthetic, warm dark palettes, and expressive motion. Use when user mentions ASCII art, dot matrix, scanline effects, terminal aesthetic, pixelated figures, block letterforms, "vibes", "make it vibes", "use the aesthetic from frontend-vibes", warm blacks, terracotta accents, Anthropic-adjacent design, technical romance, or wants emotionally expressive UI that balances precision with organic imperfection.

Generate a self-contained HTML specimen page that renders the design system live. Triggers on: "generate specimen", "preview design system", "show design system", "specimen page", "visualize design system", or /design-systems:specimen. Produces a single HTML file with all tokens rendered as visual examples.

Audit UI code for design system compliance and AI slop markers. Triggers on: "review design system", "check design compliance", "audit UI code", "design system review", "check for AI slop", "design review", or /design-systems:review. Scans CSS, SCSS, JSX, TSX, Vue, and Svelte files against DESIGN.md rules and the AI Slop Instant Giveaways checklist.

Apply a DESIGN.md design system to the project's existing CSS/style files. Triggers on: "implement the design system", "apply design system", "set up design tokens", "apply DESIGN.md", or /design-systems:implement. Reads DESIGN.md, detects the project's CSS framework, and updates style files with design tokens.

Create an opinionated design system as a DESIGN.md file. Triggers on: "create a design system", "generate DESIGN.md", "design system for [description]", "new design system", "design system", or /design-systems:create. Generates a complete DESIGN.md with Creative North Star, Core Prohibitions, Anti-Moodboard, and full token definitions. Actively avoids AI slop patterns.

Comprehensive dependency management expertise covering TypeScript (npm, yarn, pnpm, bun), Rust (cargo), and Python (pip, poetry, uv). Use when working on projects requiring: (1) Dependency auditing and listing, (2) Checking for outdated packages, (3) Security vulnerability scanning, (4) Finding release notes and changelogs, (5) Preparing upgrade reports with breaking change analysis, (6) Inspecting transient dependencies in lock files, (7) Installing or uninstalling packages, (8) Clearing package caches, (9) Monorepo workspace dependency management, or (10) Cross-ecosystem dependency health checks.

This skill should be used when the user wants sustained divergent exploration of a theme — riffing, branching, finding unexpected angles without converging toward a vision or spec. Works independently, before ideation (to generate raw material), or after refinement (to re-expand a stripped idea). Uses a toolkit of 22 creative techniques adapted to conversational energy. Produces a branching outline of ideas explored. Triggers include "riff on", "explore ideas around", "let's brainstorm about", "what could we do with", "help me think around", "idea jam", "think out loud about", "go wild with", "free-associate on", or when the user wants to explore an idea space.

Use when creating a new git worktree. Handles gitignore setup, worktree creation, and provides guidance. Invoke with "/worktrees:new <name>" or when user mentions "create worktree", "new worktree", "start feature in worktree", or "isolated workspace".

Use for multi-subagent parallel development. A main Claude instance manages multiple concurrent subagents, each working in isolated worktrees. Invoke with "/worktrees:orchestrator <feature>" or when user mentions "orchestrate agents", "parallel subagents", "coordinate development", "multi-agent workflow", "spawn subagents", or "divide feature into tasks".

Use for independent development with PR-based integration. Multiple Claude instances work autonomously, each creating PRs to main. Invoke with "/worktrees:peer <feature>" or when user mentions "peer workflow", "independent worktree", "autonomous development", "PR-based workflow", or "multi-instance development".

This skill should be used when the user asks to "create a spec", "write a specification", "create a feature specification", "build a requirements document", "gather requirements", "develop user stories", "start discovery process", or mentions needing help with requirements analysis or product specification. Guides users through a structured discovery process where stories emerge from problem understanding rather than being predefined.

Design a writing voice through guided questioning — either from scratch or by emulating references — then generate a personalized writer skill and full writing assistant subagent. Takes ~20-25 minutes through a comprehensive questionnaire covering audience, reading level, emotional temperature, pacing, and anti-voice definition.

This skill should be used when the user asks to "create a filter", "write a tokf filter", "add a filter for <tool>", "how do I filter output", or needs guidance on tokf filter step types, templates, pipes, or placement conventions.

This skill should be used when the user wants to explore a new project idea, feature concept, or product direction that needs creative development before any design or implementation work begins. This skill should be used when the user says things like "I have an idea", "help me think through", "what should this be", "let's explore", "I'm not sure what this should be", or when requirements are vague and need shaping into a clear vision.

This skill should be used when the user wants to sharpen, distill, or pressure-test an existing idea. Triggers include "refine this idea", "what's the core of this", "help me cut this down", "is this idea any good", "challenge this idea", "help me decide if this is worth pursuing", "strip this back", "what's essential here", or when an idea has too many features and needs focus.

This skill should be used when creating Docker containers for Claude Code sandboxing, configuring isolated development environments, setting up Docker volumes and persistence, configuring port forwarding for sandboxed applications, or troubleshooting Docker container issues in sandbox contexts. Provides Docker configuration knowledge specifically for running Claude Code in contained environments.

This skill should be used when reading or writing Sandbox.toml configuration files, cloning sandbox configurations from existing projects, merging user preferences with detected settings, validating sandbox configuration compatibility, or managing sandbox metadata and settings. Provides knowledge for Sandbox.toml structure and configuration management.

This skill should be used when the user asks to "address the review", "discuss PR feedback", "plan fixes for the review", "create a plan to fix PR issues", or wants to brainstorm solutions for review findings. Use after a PR review has been posted to fetch the review comment and guide the user through creating an action plan using the brainstorming workflow.

This skill should be used when the user asks to "review this PR", "check the pull request", "review PR

Use when completing work in a worktree. Handles PR creation or direct merge, worktree removal, and branch cleanup. Invoke with "/worktrees:finish" or when user mentions "finish worktree", "clean up worktree", "merge worktree", "done with worktree", or "complete worktree work".

Quick MVP deployment to fly.io for JavaScript (Next.js, RedwoodSDK, Express), Rust (Axum, Rocket), Python (FastAPI), and generic Dockerfiles. Use when deploying applications to fly.io, setting up databases (Postgres, volumes, Tigris object storage), managing secrets, configuring custom domains, setting up GitHub Actions workflows, creating review apps for pull requests, or troubleshooting fly.io deployments. Covers complete deployment workflows from initial setup through production.

Building LLM-powered React applications with the Hashbrown library. Use when the user asks to (1) Build generative UI where LLMs render React components, (2) Add client-side tool calling for LLM-app interaction, (3) Stream LLM responses in React applications, (4) Execute LLM-generated JavaScript safely in a sandbox, (5) Build browser agents or AI-powered UIs with hashbrown, (6) Control React UI from LLM output, (7) Integrate with LLM providers like OpenAI, Anthropic, Google, Azure, Bedrock, or Ollama in React apps, (8) Create chatbots, form builders, predictive text inputs, or multi-threaded conversations with LLMs, (9) Transform natural language to structured data in TypeScript React applications.

Initialize and configure local development tooling for TypeScript, Rust, and Python projects including monorepos. Use when setting up linting (ESLint, Biome, clippy, ruff), formatting (Prettier, rustfmt, ruff), type checking (tsc, mypy), testing (Vitest, Jest, cargo test, pytest), Git hooks (lefthook for commit-msg, pre-commit, pre-push), GitHub Actions workflows, package publishing (npm, crates.io, PyPI), version management (Changesets), and automated releases. Covers both single-language projects and multi-language monorepos using Nx + pnpm workspaces.

Comprehensive code review guidance for quality, performance, and architecture across all programming languages. Use when (1) User explicitly requests code review, (2) After writing significant code changes, (3) Before commits/PRs, (4) Reviewing existing codebases, (5) Analyzing code quality, (6) Detecting performance issues, (7) Identifying architectural problems, (8) Finding code smells. Provides automated analysis scripts and manual review checklists for thorough code evaluation.

Comprehensive Python development expertise covering modern best practices, type hints, FastAPI web development, async/await, testing, and performance optimization. Use when working on Python projects requiring guidance on: (1) Modern Python features and best practices, (2) Type hints and static typing with mypy, (3) FastAPI web development, (4) Async/await and asyncio patterns, (5) Testing with pytest, (6) Data validation with Pydantic, (7) Database integration (SQLAlchemy), (8) Project structure and dependencies, (9) Performance optimization, (10) Logging and observability, or (11) Code reviews and common errors.

Create and review React components following strict TypeScript patterns, container/presenter architecture, and composition-first design. Use when asked to: (1) Create a React component, (2) Review a React component, (3) Build UI features, (4) Implement forms, (5) Add data fetching with TanStack Query. Enforces KISS/YAGNI principles, headless component patterns, Tailwind v4 styling, and Web3 security best practices.

Comprehensive React application development guidance covering project architecture, state management, performance optimization, security, and modern best practices including React 19 features. Use when (1) Working in React codebases (.jsx/.tsx files), (2) Architecting React applications (project structure, state management strategies), (3) Optimizing React performance (code splitting, memoization, virtualization), (4) Implementing security best practices (auth, XSS prevention, CSRF protection), (5) Applying React 19 features (Server Components, Actions, new hooks), (6) Setting up build configurations and tooling. Automatically triggered when working with React projects or addressing React-specific architecture, performance, or security concerns.

Comprehensive Rust development expertise covering core principles, patterns, error handling, async programming, testing, and performance optimization. Use when working on Rust projects requiring guidance on: (1) Language fundamentals (ownership, lifetimes, borrowing), (2) Architectural decisions and design patterns, (3) Web development (Axum, Actix-web, Rocket), (4) AI/LLM integration, (5) CLI/TUI applications, (6) Desktop development with Tauri, (7) Async/await and concurrency, (8) Error handling strategies, (9) Testing and benchmarking, (10) Performance optimization, (11) Logging and observability, or (12) Code reviews and best practices.

Comprehensive application security expertise covering authentication, authorization, OWASP Top 10, and security best practices. Use when (1) Implementing authentication (JWT, OAuth2, sessions, OAuth for CLI/TUI/desktop apps), (2) Adding authorization (RBAC, ABAC, RLS with Supabase/PostgreSQL), (3) Security auditing code or infrastructure, (4) Setting up security infrastructure (headers, CORS, CSP, rate limiting), (5) Managing secrets and credentials, (6) Preventing OWASP Top 10 vulnerabilities (injection, XSS, CSRF, etc.), (7) Reviewing code for security issues, (8) Configuring secure web applications in TypeScript, Python, or Rust. Automatically triggered when working with authentication/authorization systems, security reviews, or addressing security vulnerabilities.

Comprehensive TypeScript development guidance covering configuration, type safety, architectural patterns, and best practices. Use when working with TypeScript codebases for (1) TypeScript configuration and setup (tsconfig.json, strict mode), (2) Type definitions and patterns (utility types, type-safe approaches), (3) Resolving TypeScript compilation errors, (4) Applying TypeScript best practices (project structure, error handling, package selection). Automatically triggered when working in .ts/.tsx files or addressing TypeScript-specific queries.

Master Flutter's animation system with implicit and explicit animations, hero transitions, physics-based motion, and custom transitions. Use when adding animations, creating transitions, implementing hero effects, or building interactive animated experiences.