Two tools with basic schemas and descriptions, but significant gaps in parameter documentation and output schema clarity. Tool names follow verb_noun convention (getS3Object, listS3Objects), which is good. However, parameter descriptions lack detail on constraints, formats, and error recovery. Output schemas are implicit (text content only) with no documented structure for downstream tool chaining. Error handling returns plain text errors rather than structured guidance. No pagination support for listS3Objects despite potential for large result sets. Missing parameter descriptions for optional 'prefix' parameter.

• Add pagination to listS3Objects: include 'max_keys' (default 20, range 1-1000) and 'continuation_token' parameters. Return structured output with keys array, total_count, and next_token.
• Document output schemas explicitly. getS3Object should return {content: string, size_bytes: number, last_modified: string}. listS3Objects should return {keys: [{key: string, size_bytes: number, last_modified: string}], total_count: number, next_token?: string}.
• Enhance parameter descriptions: 'bucket' → 'AWS S3 bucket name (3-63 chars, lowercase alphanumeric + hyphens)'; 'prefix' → 'Optional S3 key prefix filter (e.g., "data/2024/"). If omitted, lists all keys in bucket.'
• Improve error messages with recovery hints. E.g., 'Bucket not found: verify bucket exists and you have s3:ListBucket permission. Try list_s3_objects with a known bucket first.' or 'Key not found: call list_s3_objects to discover available keys.'
• Add tool annotations (readOnlyHint: true) to both tools to signal to agents that these are safe, non-destructive operations.