Claude Code Setup

Claude Code Setup

OmniRank installs into Claude Code as a plugin, either by downloading the tagged release ZIP or by cloning the repository into the plugins directory. Once installed, Claude Code discovers two skills, audit and geo-artifacts, triggered from natural-language phrases that echo their SKILL.md description field. omnirank fix is a CLI subcommand of the same package, not a third registered skill.

Verified against .claude-plugin/plugin.json, scripts/build-skill-zip.sh, and the two SKILL.md files.

Path 1: download the packaged skill ZIP

curl -LO https://github.com/bemoshiur/OmniRank/releases/latest/download/omnirank-skill.zip
unzip omnirank-skill.zip -d ~/.claude/plugins/

This is the README's documented install path. The release ZIP is built by scripts/build-skill-zip.sh and published by .github/workflows/release.yml on every v* tag push. Unzipping it stages a top-level omnirank/ directory, so the command above leaves you with ~/.claude/plugins/omnirank/. Contents mirror the archive listing for the current release — .claude-plugin/plugin.json, both skill directories under skills/, the Python package (including fixes/, registry.py, locator.py, and framework.py, all new since v0.2.1), the JSON Schemas, and the config template — everything needed to run the CLI and have both skills discoverable, without cloning the full repository (Node sources, tests, and CI config are intentionally left out).

Build the archive yourself instead of downloading it:

./scripts/build-skill-zip.sh          # -> dist/omnirank-skill-<version>.zip

Path 2: clone the repository

git clone https://github.com/bemoshiur/OmniRank.git ~/.claude/plugins/omnirank

This gets you the full repository, including tests, CI configuration, and the Node generator source — useful if you intend to contribute, or want the Node in-repo GEO path available (see GEO-Artifacts-Skill#node--in-repo-path). It also means you can git pull to update rather than re-downloading a new release archive.

Either path leaves plugin.json declaring "skills": "./skills/" — Claude Code loads skills from that relative path.

How do I confirm the skills registered?

Check the files are actually present at the path the manifest declares:

ls ~/.claude/plugins/omnirank/.claude-plugin/plugin.json
ls ~/.claude/plugins/omnirank/skills/audit/SKILL.md
ls ~/.claude/plugins/omnirank/skills/geo-artifacts/SKILL.md

All three must exist. If .claude-plugin/plugin.json is missing, the ZIP was unzipped to the wrong location or the clone did not complete; if a SKILL.md is missing, re-download or re-clone. plugin.json's "version" field reads "0.4.0" on a current checkout.

Confirm each skill's identity by reading its frontmatter (what Claude Code actually parses to register the skill and decide when to trigger it):

head -4 ~/.claude/plugins/omnirank/skills/audit/SKILL.md
head -4 ~/.claude/plugins/omnirank/skills/geo-artifacts/SKILL.md

Expected output — the real frontmatter from both files:

---
name: audit
description: Use when asked to audit a site's SEO, check AEO or answer-engine readiness, diagnose why a page is not ranking or not being cited by AI, verify structured data, or run pre-deploy discoverability checks on built HTML.
---

---
name: geo-artifacts
description: Use when asked to generate or fix llms.txt, llms-full.txt or facts.json, make a site citable or ingestible by ChatGPT, Claude, Perplexity or Gemini, or publish machine-readable ground truth for AI crawlers.
---

The behavioural check: open a Claude Code session in a project directory and use one of the trigger phrases below. If the skill registered, Claude follows the audit or geo-artifacts workflow (reading SKILL.md, running the CLI, citing gate ids) rather than answering from general knowledge alone.

What should I say to trigger each skill?

Claude Code matches a skill against its description frontmatter, not a fixed command name — these phrases work because they echo the actual clauses in each skill's description above, not because they are hardcoded trigger strings.

To trigger audit

• "Audit https://example.com for SEO issues."
• "Check AEO / answer-engine readiness for this site."
• "Why isn't this page ranking?" / "Why doesn't ChatGPT cite us?"
• "Verify our structured data / JSON-LD."
• "Run a pre-deploy discoverability check on the built HTML before we ship."

To trigger geo-artifacts

• "Generate llms.txt for this site."
• "Fix llms-full.txt" / "Fix facts.json."
• "Make our site citable by ChatGPT, Claude, Perplexity, or Gemini."
• "Publish machine-readable ground truth for AI crawlers."

To use omnirank fix

There is no SKILL.md description to match against — fix is not a Claude Code skill, it is a CLI subcommand of the same package the audit skill already installs. Run it directly, from inside a Claude Code session's terminal access or your own shell:

python3 -m omnirank.cli fix https://example.com --root .

See Fix-Preview for the full flag reference and what its output looks like.

What will not trigger either skill?

Both SKILL.md files list explicit "When NOT to use" cases. Asking Claude to "write our JSON-LD" or "emit schema for this entity type" will not trigger either shipped skill — that is aeo-onpage, which is on the roadmap and not built yet. Asking it to "submit this URL to Google" will not trigger anything either — that is the unshipped indexing skill. Asking it to "apply the fix" or "write the canonical tag for me" will also not do anything by itself: omnirank fix only prints a diff, and no shipped skill applies an edit to your source tree. See Roadmap for the full list of what does not exist yet.

See also

• Quick-Start — installing the CLI itself (not the plugin) for local/CI use
• Audit-Skill and GEO-Artifacts-Skill — what each skill actually does once triggered
• Fix-Preview — what omnirank fix does and why it has no Claude Code skill of its own
• Roadmap — which skills are planned versus shipped