Server has 27 tools with complete input schemas derived from Zod validators and descriptions present for all tools. Naming follows verb_noun convention consistently (create_, get_, list_, search_, update_, delete_). However, parameter descriptions are minimal (often 10-30 chars), lacking actionable constraints, format guidance, and dependency hints. Output schemas are not documented, LLMs cannot predict response structure. Error handling is categorized (GitHubError subclasses) but recovery guidance is absent from tool descriptions. No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite clear risk levels (WRITE, IRREVERSIBLE, DESTRUCTIVE). Destructive operations (delete_repository, merge_pull_request) lack confirmation or dry-run patterns.

• Expand parameter descriptions to 50-100 chars and include format/constraint guidance. E.g., 'Branch name (alphanumeric, hyphens, slashes; must not exist if creating new branch)' instead of 'Branch to create/update the file in'.
• Document output schemas in tool descriptions or as a separate schema field. E.g., 'Returns: {repositories: [{id, name, url, owner, private, description}], total_count, page}' so LLMs know what to extract.
• Add tool annotations: mark delete_repository and merge_pull_request with destructiveHint=true; mark read-only tools with readOnlyHint=true; mark idempotent tools with idempotentHint=true.
• Implement confirmation pattern for irreversible operations: add a 'confirm' boolean parameter (default false) to delete_repository and merge_pull_request. Require explicit confirmation before executing.
• Enhance error messages with recovery guidance. E.g., 'Repository not found. Try search_repositories() with owner and partial name, or verify the owner/repo parameters are correct.'