Claude Skill

geoskills-cli

Agent-native command surface for discovering, inspecting, and validating the geo-skills toolkit. Use when an operator or coding agent needs to run geoskills doctor, enumerate bundled skills or prompts, validate the full repository contract, validate an owner-intent handoff for To

LLM Mart · 0 points · 17 views 0 listing impressions 0 install-command copies
Virus-scanned Reviewed automatically before listing.

Full trust report

Download thesmokedev-geo-skills-skills_geoskills-cli-1d09807.zip · 1 KB
Part of thesmokedev/geo-skills — 34 skills

Install

skills CLI npx skills add https://github.com/TheSmokeDev/geo-skills/tree/main/skills/geoskills-cli
Claude Code claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install thesmokedev-geo-skills@llmmart
Git git clone https://github.com/TheSmokeDev/geo-skills.git

The skills CLI installs just this skill, for any of its supported agents. Claude Code installs the whole thesmokedev/geo-skills collection as a plugin from our marketplace. Git is the plain clone.

Skill manifest

GEO Skills CLI

Use the installed geoskills command before reaching into bundled scripts directly.

Start

geoskills --json doctor
geoskills --json validate
geoskills skills list --json
geoskills prompts list --json

Use --json for agent consumption. Omit it for human-readable output.

Commands

  • geoskills skills show <name>: print one canonical SKILL.md.
  • geoskills prompts show <name>: print one paste-ready prompt.
  • geoskills owner-map validate <file> --json: fail closed on an invalid GEO-to-TokenMax handoff.
  • geoskills inspect <url> --mode page|robots|llms|sitemap|blocks|full --json: run the deterministic read-only web inspector. Install the web extra first.
  • geoskills citability <url> --json: score extracted page passages. Install the web extra first.

Rules

  • Inspect before proposing changes.
  • Preserve the distinction between observed data, calculated values, inference, and proposed action.
  • Treat doctor, validate, discovery, owner-map validation, inspection, and citability as local/read-only commands.
  • Do not represent a deterministic inspection as a full LLM audit.
  • Do not deploy, publish, submit URLs, post to communities, mutate provider settings, or spend against a connected API from this skill.
  • On nonzero exit, read the JSON error or stderr and repair the contract rather than bypassing the gate.
Files (geo-skills)
  • agents
    • openai.yaml 214 B
      interface:
        display_name: "GEO Skills CLI"
        short_description: "Run and validate the GEO toolkit CLI"
        default_prompt: "Use $geoskills-cli to inspect, validate, or run the installed GEO Skills command surface."
      
  • SKILL.md 1.7 KB
    ---
    name: geoskills-cli
    description: Agent-native command surface for discovering, inspecting, and validating the geo-skills toolkit. Use when an operator or coding agent needs to run geoskills doctor, enumerate bundled skills or prompts, validate the full repository contract, validate an owner-intent handoff for TokenMax, or run the deterministic read-only page inspector and citability scorer.
    ---
    
    # GEO Skills CLI
    
    Use the installed `geoskills` command before reaching into bundled scripts directly.
    
    ## Start
    
    ```bash
    geoskills --json doctor
    geoskills --json validate
    geoskills skills list --json
    geoskills prompts list --json
    ```
    
    Use `--json` for agent consumption. Omit it for human-readable output.
    
    ## Commands
    
    - `geoskills skills show <name>`: print one canonical `SKILL.md`.
    - `geoskills prompts show <name>`: print one paste-ready prompt.
    - `geoskills owner-map validate <file> --json`: fail closed on an invalid GEO-to-TokenMax handoff.
    - `geoskills inspect <url> --mode page|robots|llms|sitemap|blocks|full --json`: run the deterministic read-only web inspector. Install the `web` extra first.
    - `geoskills citability <url> --json`: score extracted page passages. Install the `web` extra first.
    
    ## Rules
    
    - Inspect before proposing changes.
    - Preserve the distinction between observed data, calculated values, inference, and proposed action.
    - Treat `doctor`, `validate`, discovery, owner-map validation, inspection, and citability as local/read-only commands.
    - Do not represent a deterministic inspection as a full LLM audit.
    - Do not deploy, publish, submit URLs, post to communities, mutate provider settings, or spend against a connected API from this skill.
    - On nonzero exit, read the JSON error or stderr and repair the contract rather than bypassing the gate.
    

Comments (0)

Sign in to join the conversation.

No comments yet.

Reviews (0)

No reviews yet.

Related