HiveTrade MCP has 8 tools with complete input schemas and descriptions. Naming follows verb_noun convention (trade_*, hive_earn_*). Descriptions are substantive (100-250 chars), explaining WHAT and WHEN. However, parameter descriptions lack actionable constraints (no ranges, patterns, or format guidance). Output schemas are undocumented, LLMs cannot predict response structure. Error handling is minimal; tools return JSON but lack recovery guidance. Security is reasonable (no exposed secrets in params), but no permission gates or audit trail declarations. Composition is sound: tools are single-responsibility and chainable (e.g., create_invoice → get_invoice). Missing: per-parameter type validation hints, output schema documentation, and error classification.

• Document output schemas for all tools. E.g., trade_get_fees should specify: { fees: { hive_fee_pct: number, swift_fee_usd: number, fx_spread_pct: number, settlement_time_hours: number }, ... }. This enables LLMs to chain calls and extract the right fields.
• Add constraint hints to parameter descriptions. E.g., 'amount_usd (number, 5000 - 250000): Invoice amount in USD. Minimum $5K for SMB tier, maximum $250K for MVP.' Use enums for settle_currency and settle_chain instead of free-form strings.
• Implement error classification in responses. Return { ok: false, error_code: 'INVOICE_NOT_FOUND', message: 'Invoice XYZ not found. Check invoice_id format or call trade_get_listing() to discover valid invoices.', retryable: false }. This guides agent recovery.
• Add permission declarations to tool descriptions. E.g., 'Requires: write:invoice, read:buyer_profile. Scope: cross-border trade finance.' This enables least-privilege agent configuration.
• For trade_create_invoice and trade_dispute_invoice, add a dry_run parameter (boolean, default false) to preview outcomes without committing. E.g., dry_run=true returns estimated fees and settlement time without creating the invoice.