Suede Creator Skills: 71 Open-Source Agent Skills Released

What changed

Jason Colapietro has launched Suede Creator Skills, an open-source project offering 71 distinct agent skills designed for use with AI models like Claude Code and OpenAI Codex. The collection aims to provide founders, marketers, and small teams with reusable workflows for tasks spanning search visibility, content creation, and business operations. Key areas addressed include SEO audits, content planning for AI search, product marketing, and business process assessment. The skills are distributed as public, broadly reusable skill folders, with individual skills/ /SKILL.md files for inspection. The project is MIT licensed, with adapted components retaining their upstream notices. Notably, the release emphasizes a commitment to privacy with no binaries, telemetry, or accounts required.

Installation is streamlined across multiple platforms. For Claude Code, users can install the full pack via /plugin install suede-skills@suede or opt for focused subsets like suede-marketing (42 skills) or suede-code (review, grade, ship-gate). Codex users can install the complete plugin using codex plugin add suede-skills@suede-codex. The Hermes Agent (Nous Research) supports subscribing to the repository as a skills tap or installing individual skills. For broader compatibility, a skills CLI command npx skills add JasonColapietro/suede-creator-skills is available for any agent. Additionally, a git clone and bash install.sh script provides a local installation of all skills.

Advanced orchestration skills are also included. suede-graph-flo-xr orchestrates a Suede Thought Graph shipping search, involving plan generation, scoring, adversarial refutation, and aggregation, with call caps for different effort levels. suede-agent-teams coordinates complex changes across agent lanes, incorporating collision detection, RFC modes, and handoff checklists. suede-ship-copy is a dedicated orchestration DAG for high-stakes writing tasks, featuring multiple research lenses, claim audits, and adversarial refutation.

The suede-code skill offers an A-F grading system for code quality, assessing correctness, security, data handling, domain truth, UX, tests, and deploy readiness. It includes instant F-grade triggers for critical issues like hardcoded secrets.

Why it matters for builders

This extensive collection of agent skills empowers builders by providing ready-to-use components for a wide array of AI-driven tasks. Developers can integrate these skills into their own agent frameworks or leverage them directly within supported environments like Claude Code and Codex. The open-source nature encourages customization and contribution, allowing builders to adapt the skills to specific project needs or even extend their functionality. This significantly lowers the barrier to entry for implementing complex agent behaviors and sophisticated workflows.

Practical impact

Builders can immediately leverage Suede Creator Skills for tasks such as auditing web pages for search visibility using suede-seo-audit, planning content for AI search with suede-ai-seo, or defining product positioning with suede-product-marketing. For code quality, the suede-code skill can provide an A-F grade, identifying potential issues before deployment. The project's clear installation instructions for various platforms mean developers can quickly integrate these skills into their preferred AI development environment. Experimenting with the orchestration skills like suede-graph-flo-xr can offer insights into building more robust and complex agentic workflows.

Caveats and source limits

The provided source material details the capabilities and installation methods of Suede Creator Skills but does not include independent benchmark results or performance metrics for the skills themselves. Specific pricing information is not available as the project is open-source. While the project mentions MIT licensing for original work, details on upstream notices for adapted components are general. The source also notes that some advanced workflows, like suede-graph-flo-xr, have specific execution requirements and call caps, but the exact impact of these caps on task completion time or quality is not quantified. The availability of these skills within specific AI model marketplaces or their compatibility with future versions of Claude Code or Codex beyond the current release is not specified.