Server has 5 tools with complete input schemas and descriptions present. However, descriptions are generic and lack actionable context for LLM selection. Parameter descriptions are minimal (e.g., 'ID or key of the issue' without format guidance). Output schemas are not documented, LLMs cannot predict response structure for chaining. No error handling guidance. Tool names follow verb_noun convention (good), but descriptions do not explain WHEN to use each tool or dependencies between them. Schemas are well-formed JSON Schema with types, but lack constraints (enums, patterns, ranges) that would prevent invalid LLM inputs.

• Document output schemas for all tools. E.g., 'Returns: {issues: [{key: string, summary: string, status: string, ...}], total: number, nextPageToken?: string}'. This enables LLMs to chain tools correctly.
• Expand descriptions to 100 - 200 chars with WHEN/WHY context. E.g., 'Search for issues matching a JQL query. Use this to find issues by status, assignee, or custom criteria. Returns up to 50 results; use nextPageToken for pagination.'
• Add constraints to parameters: maxResults as integer with min=1, max=100; jql as string with pattern hint; fields as enum of valid Jira field names (summary, description, status, assignee, etc.).
• Add error recovery guidance in descriptions. E.g., 'If project key is invalid, call list_projects() to see available keys.' This guides LLM self-correction.
• Mark create_issue and update_issue with destructive/state-modifying indicators in descriptions. E.g., 'Creates a new issue (irreversible). Requires valid project key and issue type.'