AI Tools for Creators
conorbronsdon/ai-tools-for-creators
Skills e MCP servers prontos para criadores potencializarem agentes de IA
Ver no GitHubColeção curada de skills, MCP servers e ferramentas de workflow para criadores de conteúdo e profissionais do conhecimento. Inclui instruções drop-in para agentes de IA (Claude Code, Cursor, Windsurf) sem necessidade de instalação de código, além de ferramentas para gestão de sessão, detecção de padrões de escrita IA e auditoria de consistência.
Instalação
npx convez-ai conorbronsdon/ai-tools-for-creators1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
Casos de uso
- Criar e auditar conteúdo com detecção de padrões de escrita IA
- Gerenciar contexto persistente entre sessões de trabalho com agentes
- Auditar inconsistências e drift de dados em documentação
- Gerar novas skills automaticamente a partir de descrições em linguagem natural
- Recuperar e reconectar worktrees órfãs após falhas de agentes
Veredito

Repositório bem organizado com coleção genuína de skills testadas pelo autor. Foco em criadores de conteúdo é claro, mas valor maior está nas ferramentas de gerenciamento de contexto e auditoria — aplicáveis além de escrita. Comunidade ativa (podcast Chain of Thought) agrega credibilidade.
README
A tradução para o português deste README ainda está na fila.
AI Tools for Creators
A curated collection of AI skills, MCP servers, and workflow tools for content creators.
This is a living collection. Some tools are mine, most are things I've found and vouch for. If something's here, I've used it or someone I trust has.
Skills
What's a skill? A markdown file that teaches an AI coding agent how to do a specific task. Drop it into your project, and the agent follows the instructions. No code to install — just a file.
Drop-in instruction files for AI coding agents (Claude Code, Cursor, Windsurf, etc.). Copy the file, point your agent at it, done.
Writing & Content
| Skill | What it does | Author |
|---|---|---|
| avoid-ai-writing | Audit & rewrite content to remove AI writing patterns — 49 pattern categories + a 109-word replacement table. Rewrite / detect / edit-in-place modes, optional voice profiles (casual/professional/technical/warm/blunt), and a zero-dependency detector engine. Installs as a Claude Code / Cowork plugin or a standalone skill. | @conorbronsdon |
Productivity & Workflow
| Skill | What it does | Author |
|---|---|---|
| session-management | Four commands (/start, /end, /update, /today) that give Claude Code memory across sessions. State files, session logs, daily heartbeats. | @conorbronsdon |
| reconcile | Tripwire check for multi-session drift — scans commits and state files for inconsistencies from parallel sessions. | @conorbronsdon |
| recover | Scan for orphaned worktrees and stale branches after crashes. Read-only by default, cleanup requires approval. | @conorbronsdon |
| skill-creator | Generate new skills from plain-language descriptions. Scaffolds SKILL.md, command file, and CLAUDE.md additions. | @conorbronsdon |
| ssot-check | Single-source-of-truth drift auditor. Finds facts hand-copied across docs (episode counts, prices, stats), builds a manifest of canonical locations, and flags every copy that drifted. | @conorbronsdon |
| agent-memory-kit | The curation loop for agent memory — capture, recall, and a read-only curator that finds rot and contradictions before your agent is confidently wrong. Plain markdown and JSON, five slash commands, human-reviewed diffs. | @conorbronsdon |
Development & Code Review
| Skill | What it does | Author |
|---|---|---|
| code-review | Multi-agent PR review — orchestrates Copilot + parallel subagents (adversarial, operational, reference-comparison) sized to PR risk. Catches architectural P0s that single-pass review misses. | @conorbronsdon |
| eval-integrity | Audit an LLM benchmark repo for credibility practices — pre-registration, contamination guards, holdout hygiene, judge validity, statistical honesty. Answers: would your published numbers survive an adversarial reviewer? | @conorbronsdon |
Research & Booking
| Skill | What it does | Author |
|---|---|---|
| angel-diligence | Pre-investment research and deal-memo generation. Parallel web research with strict citation rules; separates verified facts from claims; ends in a verdict scaffold, never an invest/pass call. | @conorbronsdon |
| guest-circuit | Map a prospective podcast guest's appearance circuit — where they've been, their stump speech, and the unclaimed angle for your show. Uses Podcast Index when configured, web search otherwise. | @conorbronsdon |
MCP Servers
What's MCP? Model Context Protocol — a standard way to connect AI agents to external tools and data (calendars, databases, APIs). Think of it as a USB-C port for AI.
MCP servers give AI agents access to real tools and data.
Content & Publishing
| Server | What it does | Author |
|---|---|---|
| Transistor-MCP | Full Transistor.fm API access — episodes, analytics, transcripts, show management. | @conorbronsdon (fork of @gxjansen) |
| substack-mcp | Read posts, manage drafts on Substack. No publish or delete by design — safe for agent workflows. | @conorbronsdon |
| podcastindex-mcp | Podcast Index API — search by person/topic, trending podcasts, feed health checks, cross-platform episode discovery. | @conorbronsdon |
| op3-mcp | OP3 (Open Podcast Prefix Project) analytics — downloads over time, listener geography, app share, per-episode breakdowns. Read-only by design. | @conorbronsdon |
Productivity & Workspace
| Server | What it does | Author |
|---|---|---|
| gws-mcp-server | Google Workspace access — 41 tools across Gmail, Calendar, Drive, Sheets, Docs, and Tasks. Built on the official gws CLI. | @conorbronsdon |
Benchmarks & Evaluation
If you're building with agents, you eventually need to measure them. Tools for that:
| Project | What it does | Author |
|---|---|---|
| cot-bench | Open agent evaluation leaderboard. Three judges score every scenario (two open-weight, one frontier reference) across CLEAR-aligned metrics: efficacy, cost, reliability, latency. Rubrics are code, and every raw judge score is published so you can audit the results. | @conorbronsdon |
| podcast-benchmark | Benchmark any podcast against its peers using only public data — catalog depth, cadence, transcript availability, feed hygiene. No download estimates, every number sourced and timestamped. | @conorbronsdon |
Web Apps
Standalone tools that run in the browser. No agent, no install.
| App | What it does | Author |
|---|---|---|
| track-finder | Paste a tracklist or a YouTube playlist URL and get search links for every track on SoundCloud, Spotify, YouTube, Apple Music, or Beatport. Progress saves in your browser. Built for DJs and playlist builders. | @conorbronsdon |
How to Use These
Skills (Claude Code)
Most skills follow the agentskills.io standard and work across tools. For Claude Code specifically:
- Download the skill file (usually
SKILL.mdor similar) - Place it in your project directory or reference it in your CLAUDE.md
- Invoke it — most skills include slash command instructions
Skills (Other Agents)
Skills in the agentskills.io format work with Cursor, Windsurf, Cline, OpenHands, and 40+ other tools. Check each skill's README for tool-specific setup.
MCP Servers
Each MCP server has its own install instructions. Generally:
- Clone the repo
npm install && npm run build- Add the server to your Claude Code or Claude Desktop config
See Anthropic's MCP documentation for setup details.
Contributing
This is a curated list, not a dump. If you've built or found a tool that's genuinely useful for content creators, knowledge workers, or anyone who ships ideas for a living:
- Open an issue describing the tool and why it's worth including
- Or submit a PR adding it to the right section
Quality bar: you've actually used it, not just bookmarked it.
About
Maintained by Conor Bronsdon. I host the Chain of Thought podcast covering AI infrastructure, developer tools, and how practitioners actually use this stuff. Many of these tools were built or discovered through conversations with guests on the show.
New to AI tools? Check out AI Learning Resources — a curated path from "what is AI?" to building your own workflows.
If you want to go deeper on any of these topics, the podcast is a good place to start.
Disclaimer
All views, opinions, and statements expressed on this account are solely my own and are made in my personal capacity. They do not reflect, and should not be construed as reflecting, the views, positions, or policies of Modular. This account is not affiliated with, authorized by, or endorsed by Modular in any way.
License
MIT (see LICENSE).