Suede Skills MCP
Read-only access to 71 Suede skills: discovery, install options, SEO audits, A-F grading.
Runs on your machine from its package; the steps are below.
What it does
- Reads
- Writes
- High impact
At most: only reads or looks up information.
What it costs
FreeOdel charges nothing per call. The publisher may charge for its own service.
Runs locally over stdio
This server isn't hosted — your MCP client launches it from a package registry. Use one of the commands below, or drop the config into your client (e.g. Claude Desktop).
Run
npx -y suede-skills-mcp
MCP client config
{
"mcpServers": {
"suede-skills-mcp": {
"command": "npx",
"args": [
"-y",
"suede-skills-mcp"
]
}
}
}
Install · The skills · Docs site · Skill catalog · Blog · MCP · @johnnysuede
By Jason Colapietro / Suede AI
Suede Creator Skills for SEO and business growth
Suede Creator Skills gives founders, marketers, and small teams reusable workflows for SEO audits, AI search visibility, conversion copy, and business operations. Run them in Claude Code or OpenAI Codex, with instructions you can inspect and adapt to your business.
Start with the business task:
- Search visibility: audit a page with
suede-seo-audit, or plan content for AI search withsuede-ai-seo. - Positioning and conversion: define your audience and offer with
suede-product-marketing, then write page copy withsuede-copy. - Business operations: map bottlenecks with
suede-ops-assessmentand define the system withsuede-ops-architecture.
These are public, broadly reusable, open-source skill folders. Read each skills/<name>/SKILL.md before you install it. The pack also covers campaign planning, code review, app delivery, and creator-rights work. Original work is MIT licensed; adapted components carry their upstream notices beside the source. No binaries, no telemetry, no accounts.
Install in 30 seconds
Claude Code — add the marketplace, install the pack:
/plugin marketplace add JasonColapietro/suede-creator-skills
/plugin install suede-skills@suede
suede-skills installs every skill. Want less? Three focused subsets: /plugin install suede-marketing@suede (43 marketing and growth skills), /plugin install suede-agent-workflows@suede (Suede Thought Graph shipping, orchestration, workflows, evals), and /plugin install suede-code@suede (review, grade, ship-gate).
Codex — add the Codex-native marketplace, install the complete plugin:
codex plugin marketplace add JasonColapietro/suede-creator-skills --ref main
codex plugin add suede-skills@suede-codex
The Codex plugin loads every skill and registers 1 read-only MCP server with 9 tools. Restart Codex after installing or updating.
Hermes Agent (Nous Research) — this repo is a skills tap. Subscribe to browse and search the whole pack, or install one skill without subscribing:
hermes skills tap add JasonColapietro/suede-creator-skills
hermes skills search seo
hermes skills install JasonColapietro/suede-creator-skills/skills/suede-seo-audit
Hermes installs the whole skill folder — references/, agents/, evals/ — into ~/.hermes/skills/, security-scans it first, and shows it as /suede-seo-audit in chat. It carries the skills only: the MCP discovery server and the bundled Suede Thought Graph workflow are Claude Code and Codex surfaces.
Any agent (Cursor, Copilot, Windsurf, Claude Code, Codex) via the skills CLI:
npx skills add JasonColapietro/suede-creator-skills
Prefer a clone? install.sh copies every skill into ~/.claude/skills/ and prints the installed count:
git clone https://github.com/JasonColapietro/suede-creator-skills.git && bash suede-creator-skills/install.sh
You can also copy individual skill folders into .claude/skills/ (project) or
~/.claude/skills/ (user). One exception: the hardened suede-graph-flo-xr JavaScript
workflow also requires the skill's bundled agents/suede-graph-flo-xr-*.md profiles in
~/.claude/agents and macOS sandbox-exec. The suede-skills and
suede-agent-workflows Claude plugins, plus install.sh, install those profiles;
a skill-folder-only or generic skills-CLI install does
not. Codex can use the skill's orchestration contract, but does not execute this
Claude Workflow runner. The caller passes the namespace explicitly on every
launch: suede-skills for the full plugin, suede-agent-workflows for the
focused plugin, and the empty string for clone or manual user-agent installs.
More install routes (single-skill Codex installer, project-level copy, MCP)
Codex — install one skill from GitHub:
python3 ~/.codex/skills/.system/skill-installer/scripts/install-skill-from-github.py \
--repo JasonColapietro/suede-creator-skills \
--path skills/suede-agent-teams
Swap skills/<skill> for any skill folder, or pass extra skills/<name> paths to install several at once. Restart Codex after installing.
Claude Code — project-level copy of a single skill:
git clone https://github.com/JasonColapietro/suede-creator-skills.git /tmp/suede-creator-skills
mkdir -p .claude/skills
cp -R /tmp/suede-creator-skills/skills/suede-agent-teams .claude/skills/
For a user-level install, copy into ~/.claude/skills/ instead. Claude.ai and organization-managed Claude Skills may use upload or admin-managed flows instead of filesystem copy. Review skill contents before enabling code execution.
MCP — dependency-free stdio server:
node mcp/suede-skills-mcp.mjs --profile all
Exposes 9 tools (list_suede_skills, list_suede_specialties, search_suede_skills, get_suede_skill, suede_install_options, suede_copy_seo_audit, suede_visibility_grade, suede_code_grade, suede_qa_checklist), 7 resources, and 5 prompts over JSON-RPC.
Try it right now
Install the pack, then give your agent a page to audit:
Use suede-seo-audit to audit this page for search visibility and return prioritized findings with exact copy fixes: [your page URL]
For a nontrivial multi-file change, run the ship DAG:
Use suede-graph-flo-xr to take this change through the canonical Suede DAG: research it, build it in disjoint lanes, refute the review findings, and hand back evidence.
If the pack saves you an hour, star the repo — stars are how other builders find it.
Orchestration: broad outcomes, adversarial review
Tell your agent "max effort, spare no compute, fix everything" and most of the time you get enthusiasm, not engineering. The orchestration lane turns that intent into structure:
suede-graph-flo-xrruns a Suede Thought Graph shipping search on a repo: generate competing plans, score and prune them, adversarially refute and improve survivors, aggregate compatible lanes, then build, review, and gate only the selected plan. Light/standard/deep are capped at 55/110/200 calls. Production reads only; it never deploys.suede-agent-teamswires complex changes into coordinated agent lanes with WIP collision detection, RFC mode, feature-flag strategy, rollback trees, and a handoff checklist that won't close without evidence. Its public-contribution mode adds scored issue queues, atomic leases, isolated worktrees, and authority-gated contribution packets.suede-ship-copyis a separate copy-only orchestration DAG for one high-stakes piece of writing: five blind research lenses, a claim audit that closes the set of assertable facts, disjoint section writers, adversarial refutation, a deslop pass, and a publish-readiness gate.
The A–F ship grade
suede-code runs a findings pass (TypeScript, React, Next.js, OWASP, and database checklists), then returns a grade card scoring seven lanes: Correctness, Security and permissions, Data and state, Domain truth, UX and release behavior, Tests and verification, and Deploy readiness. The weakest lane caps the grade, and auth and payment surfaces carry grade caps of their own.
If it hits an instant-F trigger — a hardcoded secret, a permission check bypassable via a request param — the grade locks to F with the exact file and line, and no other lane can raise it. A polite review tells you what could be better. A grade tells you whether to ship.
The skills
Every link below goes to the skill's folder in this repo; the skill catalog has the same list with install commands per skill.
Agent orchestration & workflows
| Skill | What it does |
|---|---|
suede-agent-teams | Coordinate agent lanes and public contribution programs with collision checks, atomic issue leases, review gates, and a signed handoff |
suede-graph-flo-xr | Suede Thought Graph repo shipping: search competing plans, score and prune, refute and improve, aggregate compatible lanes, then mutate only the selected plan; 55/110/200-call caps; production reads only, never deploys |
suede-ship-copy | Copy-only orchestration DAG for one high-stakes piece: blind research, claim audit, section writers, refutation, publish gate |
suede-ai-eval | AI-SPEC artifacts, failure-mode rubrics, eval cases, and acceptance gates for LLM, RAG, classifier, and agent surfaces |
suede-recommend-next-action | Scores candidate moves on goal fit, unblocking, evidence, urgency, and leverage, then hands back one recommendation as a short runnable prompt |
suede-workflow-skills | Umbrella skill that loads the full pack |
Code quality & shipping
| Skill | What it does |
|---|---|
suede-code | Review + A–F grade in one pass |
suede-code-review | Deep findings with Accessibility and SEO lanes, no letter grade |
suede-code-grader | A–F ship verdict only, 7 lanes, instant-F triggers |
suede-ci-gate | Write CI that gates the merge — stack/lockfile detection, one required check, branch protection |
suede-clip-to-guide | Turn a clip, talk moment, or transcript into a clip-to-guide funnel package with rights routing and an evidence gate |
Design, copy & SEO
| Skill | What it does |
|---|---|
johnny-suede-design | Full design lane: brand and product surfaces, tokens, visual QA, vetted component sources with an adoption checklist |
johnny-suede-write | Full writing lane: structure, persuasion frameworks, anti-slop gate, copy score |
suede-design | Design laws, dark-mode tokens, fluid type, component rules, vetted component sources with an adoption checklist |
suede-copy | Conversion copy: headline formulas, A/B variants, anti-slop gate |
suede-deslop | Context-aware anti-slop cleanup or findings-only audit that preserves facts, voice, house style, and intentional phrasing, with a 50-point score gate |
suede-seo-audit | SEO/AEO/AI-EO, metadata, schema, internal-link, and intent audit |
suede-visibility-grader | A–F page grades for findability, clarity, CTA pull, proof, and AI citation readiness |
suede-site-alchemy | Funnel analysis, friction audit, conversion math, ranked quick-wins |
suede-launch-packaging | Package work as a launch: docs, install commands, proof links, QA |
suede-mcp-qa | QA the Suede Skills MCP before it ships |
Mobile app factories
| Skill | What it does |
|---|---|
site-to-ios-app | Convert any website into an App Store-ready iOS app: audit, 4.2 wrapper-risk gate, shell strategy, release gate |
android-app-factory | One prompt to a Play Store-ready native Android app: Kotlin + Jetpack Compose scaffold, Play Billing, ASO listing, signed release |
Consumer recovery — the negotiation lane, proven outside a repo
| Skill | What it does |
|---|---|
amazon-returns-recovery | The pack's contract negotiator, run against a real account instead of a repo: scan Amazon order/return history and digital subscriptions (Britbox, Starz, Audible, Kindle Unlimited, and more) for restocking fees, short refunds, and forgotten charges, then drive Amazon live chat to waive, refund, or cancel them — includes a validated click-path and popup workaround for the fee-dispute flow. Uses synthetic examples and requires per-item confirmation |
subscription-recovery | Audit recurring charges outside Amazon across App Store, Google Play, PayPal, bank/card evidence, and direct-bill services; report findings first and require confirmation before cancellation or refund contact |
Creator rights & release
Rights and release-prep tooling for working musicians and creators — see the skill catalog for each one: suede-campaign-in-a-box, suede-sync-packaging, suede-release-linter, suede-rights-passport, suede-rights-audit.
Marketing & growth
Forty-one skills covering paid acquisition and outbound, monetisation, lifecycle and
retention, and the measurement and operations layer — suede-ads, suede-ad-creative,
suede-cold-email, suede-prospecting, suede-public-relations, suede-directory-submissions,
suede-pricing, suede-offers, suede-paywalls, suede-signup, suede-onboarding,
suede-churn-prevention, suede-emails, suede-referrals, suede-co-marketing,
suede-community-marketing, suede-lead-magnets, suede-free-tools, suede-sms,
suede-marketing-loops, suede-analytics, suede-ab-testing, suede-revops,
suede-sales-enablement, suede-programmatic-seo, suede-content-strategy,
suede-marketing-plan, suede-marketing-psychology, suede-marketing-council, suede-instagram-growth, suede-social,
suede-aso, suede-video, suede-image, suede-product-marketing, suede-competitors,
suede-competitor-profiling, suede-customer-research, suede-marketing-ideas,
suede-ai-seo, suede-attribution.
The content-production lane adds one more workflow:
| Skill | What it does |
|---|---|
suede-newsroom | Runs one idea through a six-role content pipeline — written role contracts, an inspectable campaign record, seven distribution entryways so no two assets argue the same thing, and a keep/test/stop review |
From the blog
Essays on why the pack is built the way it is:
- Two skills, and the bugs that wrote them — Google Play publishing from the terminal, and a parity contract that keeps four surfaces of one product honest
- 71 skills installed. Your agent reads almost none of them. — progressive disclosure, Suede Thought Graph shipping search, and the MCP layer
- Why breadth is free now, and what that changes — the economics of broad skill packs
- NOT FOR: the two words that make a big skill pack work — how skills route without colliding
- Your memory file is a tax you pay on every prompt — why memory belongs in skills
- Two new lanes: a next-action recommender and a one-prompt Android app factory
Public pages
- GitHub repository
- GitHub Pages site — public documentation generated from this repo
- Skill docs catalog — every skill with install and resource links
- Installs and MCP page — install commands plus the Suede Skills MCP
- Copy bank (source) and public explainer pack
- S-Tier: the builder's book (source, PDF) — a free, roughly 29,900-word book on how Agent Skills work and how to get genuinely good with them
MCP server
The repo ships a dependency-free stdio MCP server at mcp/. It implements initialize, tools/list, tools/call, resources/list, resources/read, prompts/list, and prompts/get.
node mcp/suede-skills-mcp.mjs --profile all
| Tools (9) | Resources (7) | Prompts (5) |
|---|---|---|
list_suede_skills, list_suede_specialties, search_suede_skills, get_suede_skill, suede_install_options, suede_copy_seo_audit, suede_visibility_grade, suede_code_grade, suede_qa_checklist | catalog, specialties, plugins, copy-seo-audit, visibility-grade, code-grade, qa-checklist | discovery and audit prompts |
Safety
Core local workflows inspect current files, pages, repos, and local folders without uploading by default or granting new authority. Explicit browser, deployment, marketplace, publishing, or support workflows may use the operator's existing authenticated access only when that action is authorized. The skills do not request secrets, write to a registry, or claim legal clearance.
The creator skills generate private drafts by default — reports and packages can contain names, payment notes, file names, hashes, rights claims, and provenance notes. Review and redact before publishing or sending outside the intended workflow. A clean report is not a legal opinion, registry write, rights clearance, or payment guarantee.
CLI flags on the creator scripts: --include-hidden, --include-other, --include-absolute-paths, and --force. Scripts skip hidden files, secret-like files, dependency folders, and build outputs by default, and refuse to write a package into or under its source folder.
Every skill ships with a skill card — a release record following the NVIDIA skill-card template that states the skill's purpose, owner, license, dependencies, known risks and mitigations, output, and release evidence, so a reviewer can evaluate it without opening the source first.
Status
The creator scripts are dependency-light Python and run on the standard library. Optional enhancements use installed packages when available — Pillow for artwork dimension checks, PyYAML for YAML metadata:
python3 -m pip install PyYAML
About the creator
Jason Colapietro is the founder and CEO of Suede AI. He builds programmable IP and creator-ownership infrastructure for AI-native media. He spent years watching hired marketing firms skip the fundamentals on his own products; this pack turns those misses into reusable, inspectable agent workflows instead of one-off fixes.
Follow: X / @johnnysuede · suedeai.ai · suedeai.ai/founder
License and contributions
The original work is released under the MIT License. Adapted components retain their upstream licenses, listed in NOTICE.md.
Contributions are welcome for docs fixes, install-path corrections, lint rules, template improvements, and public-safe workflow improvements. Do not submit private catalogs, unreleased media, credentials, seed phrases, private Suede API details, payment secrets, or third-party copyrighted files.
