SEO Skills for AI Agents
Install ready-made SEO skills into Claude Code, Cursor or Codex and let the agent handle title tags, blog refreshes, backlinks and copy.
- Source repo
- Illyism/seo-skills
- Stars
- ★ 537
- Last updated
- 12d ago
- FA score
- 27/100 · Major gaps
At a glance
- Works with
- Universal · cross-platformCodex · Claude CodeChatGPT (Partial support)
- You'll need
- Typical use
- A solo developer wants a daily automated SEO pass over their own site with a fix list, using seo-autopilot.
- Main limitation
- The repo does not state which agent versions or models the skills target, and there is no evidence of cross-agent behavior testing.
- Source review
- 27/100 · Major gaps 3 safety controls not found
What does this agent do, and when should you use it?
SEO Skills is an open-source collection of SKILL.md files that gives SEO and content-marketing abilities to any agent that can read them, including Claude Code, Cursor and Codex. The repo organizes skills under four categories — seo, content, marketing and copywriting — each as <category>/<skill-name>/SKILL.md following the Agent Skills spec. You install them with npx skills add Illyism/seo-skills, either one at a time or the whole set. Coverage runs from site health checks and title-tag rewrites to post refreshes, backlink purchasing, customer-journey mapping, video transcription and landing-page copy. It is MIT licensed and maintained by Ilias Ism.
Each skill is a SKILL.md instruction file that the agent reads and follows. seo-autopilot chains SEO into a daily loop: site health checks, fixes, a keyword map, articles, comparison pages, free tools, link drafts and AI answer tracking. blitz-seo is a 30-day sprint for one money keyword. seo-title-optimization rewrites and audits title tags, H1s and meta titles. blog-refresher runs a SERP gap analysis to revive an existing post. linkdr plans a DR-tier backlink mix for a budget and produces one checkout link. On the content side, intro writes short blog intros matched to listicle, how-to or definition intent, and scribiz produces transcripts, subtitles, summaries and chapters via Scribiz plus video-to-blog workflows. Marketing skills add customer-journey, which maps a five-stage awareness journey with the Google and ChatGPT prompts people use at each stage, and diff-marketing, which turns recent commits and merged PRs into blog posts, launch posts, changelogs and emails. Copywriting covers blitz-copy's Hook → Spike → Payoff structure for landing pages, headlines, SEO titles and cold emails, plus understand-customers for pre-writing research.
- A solo developer wants a daily automated SEO pass over their own site with a fix list, using seo-autopilot.
- A content team sees one old post slipping in rankings and rebuilds it with blog-refresher's SERP gap analysis.
- A growth team planning a 30-day push on a single money keyword runs blitz-seo.
- A marketer who wants backlinks but doesn't know tiering uses linkdr to plan a DR mix within budget.
- A video-first team turns recordings into subtitles, summaries and a blog draft with scribiz.
- A copywriter producing a landing page uses blitz-copy's Hook → Spike → Payoff frame for headlines and body.
How do you install or deploy this agent?
You need Node.js and npx available locally, with network access. Install a single skill:
npx skills add Illyism/seo-skills --skill seo-autopilotInstall all skills:
npx skills add Illyism/seo-skillsList available skills:
npx skills add Illyism/seo-skills --listSkills land in the <category>/<skill-name>/SKILL.md structure defined by the Agent Skills spec.
How do you use this agent?
Open your project in an agent that reads SKILL.md files (Claude Code, Cursor, Codex, and similar), then have the agent read the relevant SKILL.md and follow its flow. To run a daily SEO pass, install and use seo-autopilot:
npx skills add Illyism/seo-skills --skill seo-autopilotFor writing tasks, install intro or blitz-copy instead:
npx skills add Illyism/seo-skills --skill blitz-copySkills such as linkdr and scribiz route through the LinkDR and Scribiz services, so expect to set up the corresponding account and budget where the skill asks for it.
What are this agent's strengths and limitations?
- SKILL.md is a portable interface, so the same skill set works across Claude Code, Cursor, Codex and other SKILL.md readers.
- Coverage spans technical SEO (titles, site health) through content, journey mapping and copywriting rather than one narrow task.
- Installation is a single npx skills add command, and you can pull one skill or the whole repository.
- diff-marketing converts commits and merged PRs directly into launch posts and changelogs, matching how developers already produce material.
- The repo does not state which agent versions or models the skills target, and there is no evidence of cross-agent behavior testing.
- linkdr requires buying managed backlinks through LinkDR and scribiz depends on Scribiz, so two skills carry vendor dependencies.
- Skill output depends entirely on the connected LLM; no model ships with the repo, so cost sits on the model side.
- The README does not document the site permissions, API keys or data-egress implications of running each skill.
How does this agent compare with similar options?
Key facts side by side with the most closely related agents.
| Agent | Source review | Form / cost | Stars | Updated | Language | Full support on |
|---|---|---|---|---|---|---|
| SEO Skills for AI Agents This agent | 27 · Major gaps | — | ★ 537 | 12d ago | — | Codex · Claude Code |
| SuperSEO Skills | 49 · Major gaps | Agent plugin / skillFreemium | ★ 360 | 1mo ago | — | Claude Code · Claude.ai |
| Marketing Skills for AI Agents | 40 · Major gaps | CLIFree | ★ 55k | 3d ago | JavaScript | Codex · Claude Code |
| LinkedIn Skills for Claude Code and Codex | 61 · Some gaps | Agent plugin / skillFree + model costs | ★ 4.9k | 3d ago | Python | Codex · Claude Code · Claude.ai |
How does FollowAgents rate this agent?
Why each dimension lost points
Only the README is visible; no SKILL.md bodies, scripts, or configs are provided, so least privilege, user confirmation, data-flow transparency, sensitive-data handling, dependency security, external effects, rollback, and source attribution cannot be confirmed from evidence. The linkdr skill explicitly describes buying managed backlinks and sending a checkout link, an external financial side effect, yet the repo shows no confirmation step, budget cap, or rollback, so external_effects scores 1; least_privilege and user_confirmation score 1 for lack of implementation evidence; sensitive_data_handling and dependency_security score 0 for no evidence; rollback scores 0 for no description; source_attribution scores 1 for the thin author/MIT note at the end of the README.
The README skill table and directory structure (<category>/<skill-name>/SKILL.md) are self-consistent, so self_consistency scores 1; installation depends on npx skills add and external sites il.ly, linkdr.com, scribiz.com whose availability is unverified, so dependency_availability scores 1; there is no description of error handling, failure messages, or fallback paths, so failure_messages scores 0.
The README names Claude Code, Cursor, Codex, and any agent reading SKILL.md, and separates SEO/content/marketing/copywriting scenarios, so audience_and_scenarios scores 2; however it does not state capability boundaries (e.g., whether linkdr auto-purchases or what seo-autopilot may touch), so capability_boundaries scores 1; triggers rely only on skill names and one-line descriptions, so trigger_precision scores 1; environment fit mentions only npx installation with no Node version, network, or offline guidance, so environment_fit scores 1.
The README organizes skills by category with a stated directory structure, so information_architecture scores 2; install commands are clear and copyable, so install_notes scores 2; skill names (seo-autopilot, blitz-seo, etc.) are consistent within the README but there is no versioning or rename policy, so naming_stability scores 1; examples_and_faq has only the skill table and one-line descriptions with no usage examples or FAQ, so it scores 1; known_limitations is entirely absent, scoring 0; the License section says MIT while metadata says unknown, an inconsistency, so license scores 1; there is no version number or changelog, so versioning_changelog scores 0; maintenance responsibility points only to the author's personal site with no issue process or SLA, so maintenance_responsibility scores 1.
The README describes output types per skill (title rewrites, SERP gap analysis, checkout links), but there are no sample outputs or quality evidence, so output_usability scores 1; the marginal-value claim over a generic agent is plausible but unverified, so marginal_value scores 1; on cost-benefit, linkdr involves real payment but there is no budget control or performance data, so cost_benefit scores 1.
All capability claims come from the README's own prose with no SKILL.md bodies, tests, or third-party corroboration, so claim_traceability scores 0; cross_source_corroboration has no independent source, scoring 0; fact_inference_separation cannot be assessed because the repo does not distinguish fact from inference, scoring 0.
- Not found in source: sensitive-data handlingUse dedicated, low-privilege, revocable API keys — never production credentials — and keep secrets out of logs.
- Not found in source: dependency securityPin versions and run a dependency audit (npm audit, pip-audit) before installing; prefer running it in a container.
- Not found in source: rollback or recovery pathBack up first, or work on a git branch or snapshot, so its changes can be undone.
- The repository provides only a README with no SKILL.md bodies, scripts, or configs, so actual skill behavior, permission scope, and safety cannot be verified.
- The linkdr skill involves real paid backlink purchases and sends a checkout link, yet no user confirmation, budget cap, or rollback is shown; review its SKILL.md manually before use.
- Installation depends on npx skills add and external sites il.ly, linkdr.com, and scribiz.com whose availability and safety are unverified.
- License metadata is unknown while the README claims MIT, an inconsistency to resolve before use.
- Publisher identity is not verified by the FollowAgents curated enterprise registry, and maintenance responsibility and update paths are unclear.