Single tool 'churnlens_scan' has a basic description (89 chars) but critically lacks parameter descriptions. The input schema defines 'target' as a string with a description ('Path to CSV ledger or Stripe export file'), but the tool description does not explain what constitutes valid input, expected file formats, error cases, or output structure. No output schema is documented. The tool name uses a verb+noun pattern (good), but the implementation is minimal and lacks the depth expected for production use. No error handling guidance, no pagination, no structured output documentation.

• Expand tool description to 150 - 250 characters. Include: (1) What it does: 'Analyzes SaaS metrics from billing data'; (2) When to use it: 'Call this to compute MRR, churn, and LTV from Stripe exports or CSV ledgers'; (3) What it returns: 'Structured JSON with monthly metrics, customer counts, and churn rates'; (4) Prerequisites: 'File must be readable and in CSV or Stripe export format'.
• Document the output schema explicitly. Example: 'Returns {"tool": "churnlens", "findings": [{"metric": string, "value": number, "month": string}], "summary": {"current_mrr": number, "churn_rate": number}}'.
• Add parameter constraints to the 'target' description: 'Absolute or relative path to a CSV file (max 100 MB) or Stripe export JSON. File must be readable by the server process. Relative paths are resolved from the server's working directory.'
• Document error cases: 'If file not found, returns error with suggestion to check path. If file is malformed, returns error with line number and expected format. If file exceeds size limit, returns error with current size and limit.'
• Add a 'format' enum parameter (optional, default='auto') to let LLMs explicitly specify 'csv' or 'stripe_export', reducing ambiguity and improving error messages.