position-me

​

Install

npx "@opendirectory.dev/skills" install position-me --target claude

npx skills add Varnan-Tech/opendirectory --skill position-me

/plugin marketplace add Varnan-Tech/opendirectory
/plugin install opendirectory-gtm-skills@opendirectory-marketplace

​

Core Capabilities

• Autonomous Navigation: Crawl the homepage, about, pricing, and blog pages.
• Visual Psychology: Take full page screenshots and analyze visual friction, layout flow, and cognitive load using the LIFT model.
• Copywriting Critique: Rewrite generic text using the PAS (Problem, Agitation, Solution) framework.
• Deep Content Audits: Read actual blog posts to identify fluff, grade technical depth, and offer new content strategies.
• Technical Readiness: Verify the presence of llms.txt, sitemap.xml, and semantic JSON-LD Schema (FAQPage, SoftwareApplication, etc) for generative search optimization.

​

Requirements

1. A multimodal LLM capable of analyzing screenshots natively.
2. Chrome DevTools MCP Server (https://github.com/ChromeDevTools/chrome-devtools-mcp).
3. Chrome CDP Skill (https://github.com/pasky/chrome-cdp-skill).
4. Terminal access to run fallback Python scripts if browser automation fails.

• Install and configure the Chrome DevTools MCP server in the agent MCP settings.
• Add the Chrome CDP skill to the agent skills directory.
• Ensure Chrome is running with remote debugging enabled (e.g., chrome --remote-debugging-port=9222).

​

Setup

​

For Developers (Manual Setup)

1. Clone or download the position-me repository.
2. Copy the position-me/ directory into the agent designated skills directory.
3. Alternatively, extract the raw SKILL.md and references/ folder into the agent custom prompt configuration.

​

For AI Agents (Automated Setup)

1. Clone or download the position-me repository into the active project workspace.
2. Read directly from the position-me/ source folder.
3. Add the SKILL.md file to the active context window or skill registry.
4. Verify access to the references/EVALUATION_SOP.md and references/REPORT_TEMPLATE.md files.

​

Execution Protocol for Agents

1. Data Gathering: Prompt the user for the target URL.
2. Connect to Browser: Use the Chrome DevTools MCP combined with the Chrome CDP skill.
3. Multi-Page Crawl: Navigate to the homepage, pricing page, about page, and at least one blog post.
4. Capture Vision Context: Take screenshots of every visited page and load them into memory.
5. Extract DOM Context: Pull the accessibility tree, search for llms.txt, and verify JSON-LD schemas.
6. Analyze: Apply the frameworks specified in references/EVALUATION_SOP.md.
7. Report: Generate a scored, highly structured teardown using exactly the format in references/REPORT_TEMPLATE.md.

​

Project Structure

position-me/
├── README.md                        # Documentation
└── position-me/                     # Raw skill source code
    ├── SKILL.md                     # Master protocol
    ├── scripts/
    │   └── extract_links.py         # Fallback crawler script
    └── references/
        ├── EVALUATION_SOP.md        # Evaluation frameworks
        └── REPORT_TEMPLATE.md       # Teardown report template