Single tool 'claudex_get_context' has a reasonable description (89 chars, within 10-1024 range) but critical schema deficiencies. Input schema is present with 'cwd' parameter typed as string with description, meeting minimum requirements. However, NO output schema is documented, LLMs cannot plan downstream calls or extract returned data. Tool name lacks action verb prefix ('get_' is present but generic for a context-retrieval tool). Parameter description is minimal (11 chars: 'The current working directory path to retrieve context for'), barely adequate. No error handling guidance, no pagination support despite likely returning large context blocks, no field-level documentation of what 'context' contains. Composition is single-purpose (acceptable), but lacks the richness expected of production tools.

• Add explicit output schema documenting all fields returned by claudex_get_context (e.g., { context: string, metadata: { lastUpdated: string, size: number }, ... }). This enables LLM planning.
• Expand parameter description to 50+ chars: 'The working directory path (absolute or relative) to retrieve rolling context for. Must be an existing directory with AGENTS.md or session history.'
• Document error cases: 'If cwd does not exist, returns error with suggestion to verify path. If no context found, returns empty context with hint to initialize session.'
• Add pagination support: accept optional 'limit' (default 50, max 500) and 'offset' parameters; return 'total_size' and 'has_more' in response to handle large contexts.
• Rename tool to 'get_rolling_context' or 'get_session_context' to clarify what 'context' means in the Claudex domain (session memory, agent state, etc.).