Findings & checks · 1 flagged
HIGH"Ignore previous instructions" command embedded in the skillSS-SKILL-INJECT-IGNORE-01 · Prompt injection · SECURITY.md
The text {match} is the classic direct prompt-injection phrasing. Placed in a skill body that the agent reads as trusted instructions, it tries to make the agent abandon its prior rules and follow whatever comes next — a full system-prompt override.
- Delete the
ignore/disregard/forget … previous instructionssentence. - If this is jailbreak-research or tutorial content, move the example into a clearly fenced, non-instruction block and label it as a quoted sample.
Trace & refs
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
SEO & GEO Skills Library
20 skills. 5 commands. Plan, audit, and monitor SEO/GEO work.
Claude Skills and Commands for Search Engine Optimization (SEO) and Generative Engine Optimization (GEO). Skill content is zero-dependency Markdown; Claude Code hooks use a small Bash runner. Content quality uses CORE-EEAT (80 items); domain trust uses CITE (40 items).
Quick Start
Install with Claude Code, any Agent Skills-compatible host, or a plain git clone:
| Tool | Install |
|---|---|
| Claude Code | /plugin marketplace add aaron-he-zhu/seo-geo-claude-skills |
| skills.sh / generic Agent Skills hosts | npx skills add aaron-he-zhu/seo-geo-claude-skills |
| Any host | git clone https://github.com/aaron-he-zhu/seo-geo-claude-skills |
Single skill: npx skills add aaron-he-zhu/seo-geo-claude-skills -s keyword-research.
If your host supports automatic skill routing, try a natural-language request:
Research keywords for my SaaS product targeting small teamsSlash-command entrypoint when your host exposes ./commands/:
/aaron:auto audit https://example.com/blog/my-articleOptional tools are documented in CONNECTORS.md; every skill also works at Tier 1 with user-provided data.
Operating Model
Every skill follows the same activation contract: Quick Start, Skill Contract, Handoff Summary, Data Sources, Instructions, Reference Materials, and Next Best Skill. Four cross-cutting skills form the protocol layer:
| Protocol skill | Role |
|---|---|
content-quality-auditor | 80-item CORE-EEAT publish gate |
domain-authority-auditor | 40-item CITE trust gate |
entity-optimizer | Canonical entity profile |
memory-management | HOT/WARM/COLD project memory (capture · promote · demote · archive · query) |
Shared refs: skill-contract.md, state-model.md, auditor-runbook.md.
Skills
<!-- SKILLS:START -->
<!-- SKILLS:END -->
Commands
Five commands, organized by SEO/GEO intent:
| Command | Use it for |
|---|---|
/aaron:auto | Describe any goal — infers intent and runs the smallest useful workflow (--deep for exhaustive/stress-test) |
/aaron:research | Keyword demand, SERP intent, competitors, content gaps, site/topic/entity maps |
/aaron:create | Brief, write, series, refresh, CMS-neutral publish package (--brief/--series/--refresh/--publish/--meta/--schema) |
/aaron:audit | On-page + CORE-EEAT quality, technical SEO, AI visibility, domain authority (--full/--tech/--visibility/--authority) |
/aaron:track | Rankings, alerts, performance reports, project memory (--alert/--report/--remember) |
Daily work normally starts with /aaron:auto, which runs the workflow implied by your goal and stops only at blocking decisions. The other four are explicit mode entrypoints.
Breaking rename note: current commands use /aaron:. Paste any old /seo:* command into /aaron:auto to recover the new route; for example, /aaron:auto /seo:audit-page https://example.com/blog/post returns /aaron:audit https://example.com/blog/post.
Command files: commands/.
Recommended Workflow
- Research:
keyword-research->competitor-analysis->content-gap-analysis. - Build:
seo-content-writer->geo-content-optimizer->meta-tags-optimizer/schema-markup-generator. - Optimize:
content-quality-auditor->on-page-seo-auditor->technical-seo-checker. - Monitor:
rank-tracker->performance-reporter->alert-manager.
For a full trust review, pair content-quality-auditor with domain-authority-auditor for a 120-item assessment. If memory-management is active, handoffs and open loops are retained in HOT/WARM/COLD memory.
References
| Reference | Purpose |
|---|---|
| core-eeat-benchmark.md | 80-item content quality benchmark |
| cite-domain-rating.md | 40-item domain authority benchmark |
| auditor-runbook.md | Auditor handoff schema, cap arithmetic, artifact gate |
| CONNECTORS.md | Optional MCP/tool connector tiers |
Contributing
See CONTRIBUTING.md. Release state is tracked in VERSIONS.md; security guidance is in SECURITY.md.
Disclaimer
These skills assist SEO/GEO workflows but do not guarantee rankings, AI citations, traffic, legal compliance, or business outcomes. Verify recommendations with qualified professionals before relying on them for major strategy or legal decisions.
License
Apache License 2.0
Star History
<a href="https://www.star-history.com/?repos=aaron-he-zhu%2Fseo-geo-claude-skills&type=date&legend=top-left"> <picture> <source media="(prefers-color-scheme: dark)" srcset="https://api.star-history.com/chart?repos=aaron-he-zhu/seo-geo-claude-skills&type=date&theme=dark&legend=top-left" /> <source media="(prefers-color-scheme: light)" srcset="https://api.star-history.com/chart?repos=aaron-he-zhu/seo-geo-claude-skills&type=date&legend=top-left" /> <img alt="Star History Chart" src="https://api.star-history.com/chart?repos=aaron-he-zhu/seo-geo-claude-skills&type=date&legend=top-left" /> </picture> </a>