Four read-only tools with basic schemas and Chinese descriptions. All tools have input schemas with proper JSON Schema structure (type, properties, required, additionalProperties). Descriptions are present but brief (20-80 chars), lacking LLM-optimized guidance on WHEN to use each tool or dependencies between them. Parameters are minimally described; 'repo' and 'ref' lack format/constraint details. No output schemas documented. Error handling exists in code (ToolError class) but not surfaced in tool definitions. No tool annotations (readOnlyHint, idempotentHint). Naming follows verb_noun pattern (github_status, github_repo_view, github_file_tree, github_file_read) but descriptions are terse and assume Chinese-speaking users.

• Expand each tool description to 50-150 characters with explicit WHAT/WHEN/WHY guidance. Example: 'github_status' → 'Check GitHub CLI authentication status and current repository. Call this first to verify gh CLI is installed and you are logged in.'
• Document output schemas for each tool. E.g., github_repo_view returns {nameWithOwner, description, visibility, defaultBranch, primaryLanguage, stargazerCount, forkCount, pushedAt}. Include field types and which fields are required for downstream tool calls.
• Add format and constraint details to parameter descriptions. E.g., 'repo: Repository in "owner/name" format (e.g., torvalds/linux)'; 'limit: Maximum results to return (1-2000, default 200)'.
• Add tool annotations to the tool definition. Mark all four tools with readOnlyHint=true and idempotentHint=true so agents know they are safe to retry and do not modify state.
• Provide English descriptions alongside or instead of Chinese. Most LLMs are trained primarily on English; Chinese descriptions may be ignored or misinterpreted by non-Chinese-trained models.