Two tools with partial schema coverage. download_attachment has a complete, well-typed input schema (7 params, all described, appropriate types/constraints). audit_pending_changes has 7 params with types and descriptions. Both descriptions are adequate (60-90 chars) but lack actionable context for LLM selection. No output schemas documented. No error handling guidance. No tool annotations (readOnlyHint, destructiveHint). Naming is clear (verb_noun pattern). No security issues detected (no secrets in params). Composition is reasonable, two distinct concerns. However, missing output documentation and error recovery patterns significantly limit production readiness.

• Add output schema documentation for download_attachment: {file_path: string, file_size_mb: number, mime_type: string, saved_locally: boolean}. For audit_pending_changes: {changes: [{target_name, risk_level, patterns_matched, clone_detected}], total_count: number, developer: string}.
• Enhance descriptions with WHEN/WHY: 'Download attachments from ServiceNow records (incidents, KB articles, etc.). Use when you need to retrieve files for analysis or local processing.' Add similar context to audit_pending_changes.
• Add tool annotations: both tools should include readOnlyHint=true in their definitions to signal to agents that they are safe to retry and have no side effects.
• Implement error recovery guidance: 'Attachment not found. Available attachments: [list]. Try audit_pending_changes to see all records.' Return structured error objects with error_code, message, and suggested_next_tool.
• Add input validation constraints to download_attachment: document max_size_mb range (1-5000), output_dir path restrictions (no traversal, max 255 chars), and filename character whitelist.