Claude Code agent
WebmasterID Agent routes work packages to Claude Code for local execution — content drafting, structured data, and technical fixes applied directly to your codebase, with your approval before anything commits or ships.
What Claude Code can do
Draft and edit content
Write title tags, meta descriptions, H1s, and body content based on work package context — keyword signals, current position, and competitor SERP analysis included.
Insert and fix structured data
Generate valid JSON-LD for Article, FAQPage, Product, BreadcrumbList, and SoftwareApplication schemas. Apply directly to your source files.
Fix technical SEO issues
Repair broken canonical tags, hreflang mismatches, missing alt text, and redirect chains — in your codebase, with changes you can review before committing.
Apply code changes
Claude Code has full access to your local file system via the MCP bridge. Changes are staged for your review — no automatic commits or deploys.
Run analysis
Query the Agent MCP server for site signals, work package details, and provider data — giving Claude the full context it needs to make informed recommendations.
The approval loop
- 01Work package queued by the Agent Brain with full signal context
- 02You open the package in the dashboard and review the finding
- 03You invoke Claude Code with the MCP bridge connected
- 04Claude reads the work package context and drafts the fix
- 05You review the proposed changes in your editor
- 06You approve — changes are committed and deployed on your schedule
Frequently asked questions
Do I need to install Claude Code separately?
Yes. The Claude Code CLI must be installed and authenticated on your machine. The WebmasterID Agent MCP bridge connects to it — it does not ship Claude Code itself.
Can Claude Code commit and push without my approval?
No. The approval gate is enforced in the Agent workflow. Claude Code can stage changes, but committing, pushing, or deploying requires explicit operator confirmation in the dashboard.
What model does the Agent use?
The Agent uses Anthropic Claude models via the official Claude API. The specific model is configurable — defaults to Claude Sonnet for cost-efficiency, upgradeable to Claude Opus for complex tasks.
Is there a cost for using Claude Code with the Agent?
Yes. Claude Code API calls are billed to your Anthropic account per token. The Agent surfaces estimated token cost before executing work packages above a configurable threshold.