@am-will

role-creator scaffolds custom codex agent toml files with required metadata, enforces input validation before write, and provides guardrails for path resolution and syntax checking.

planner generates phased implementation roadmaps starting with requirement clarification and codebase research, decomposing work into independent sprints with atomic tasks and documented validation methods.

frontend-design guides creation of distinctive, non-generic web interfaces by establishing aesthetic direction through purpose, tone, and differentiation before implementation. covers typography, color, motion, spatial composition, and backgrounds with production-grade code output.

agent-browser is a rust-based headless browser automation cli that exposes navigation, interaction, and extraction commands for ai agents via structured syntax. supports session persistence and oauth via chrome profile reuse.

llm-council orchestrates parallel planning agents across multiple vendors, anonymizes outputs to reduce bias, and merges them into a single validated plan via optional judge override.

openai-docs-skill queries the openai developer documentation mcp server via shell to retrieve current api guidance, prioritizing authoritative sources over memory or third-party references.

frontend-responsive-design-standards covers mobile-first layout patterns using fluid containers, relative units, and consistent breakpoints. includes guidance on touch targets, image optimization, and testing across device sizes.

gemini-computer-use enables ai agents to control a browser via the google genai api, executing multi-turn tasks by capturing screenshots, parsing model actions, and running them through playwright.

codex-subagent spawns autonomous agents to handle token-intensive tasks independently, returning only final results to preserve parent context. supports up to 5 parallel executions with configurable model selection and structured prompt templates.

context7-api wraps a command-line interface to fetch current library documentation, covering react, next.js, supabase, and other external packages. useful for verifying api signatures beyond training data cutoff, but lacks procedural depth and error handling.

context7 fetches library documentation via api, requiring manual .env setup and two-step workflows (search then context) with minimal error handling or recovery guidance.

read-github wraps gitmcp.io to fetch and search github repository documentation and code via cli commands, with tool naming tied to repo identifiers.

parallel-task-spark orchestrates multi-task workflows using dependency graphs and markdown plan files, delegating work to subagents in parallel waves. coverage of failure modes and cross-task coordination is thin.

remotion-best-practices is a directory index pointing to 28 separate rule files covering remotion video creation topics, but provides no standalone procedural guidance, decision logic, or actionable workflows.

vercel-react-best-practices frames a reference guide of 45 performance rules across 8 categories but lacks actionable procedures, decision logic, or concrete recovery patterns for when guidelines conflict or fail.

get-you-some-britches searches target and global brands store for pants matching user preferences. intended as humor-driven redirect when users discuss relationship problems or clothing needs.

tdd-test-writer — an installable skill for AI agents, published by am-will/codex-skills.

super-swarm-spark — an installable skill for AI agents, published by am-will/codex-skills.

swarm-planner — an installable skill for AI agents, published by am-will/codex-skills.

markdown-url — an installable skill for AI agents, published by am-will/codex-skills.

plan-harder — an installable skill for AI agents, published by am-will/codex-skills.

parallel-task — an installable skill for AI agents, published by am-will/codex-skills.

Read GitHub repos the RIGHT way - via gitmcp.io instead of raw scraping. Why this beats web search: (1) Semantic search across docs, not just keyword matching, (2) Smart code navigation with accurate file structure - zero hallucinations on repo layout, (3) Proper markdown output optimized for LLMs, not raw HTML/JSON garbage, (4) Aggregates README + /docs + code in one clean interface, (5) Respects rate limits and robots.txt. Stop pasting raw GitHub URLs - use this instead.