This is a security lab/educational server with 3 fictitious tools. All tools have descriptions and input schemas with proper JSON Schema structure. However, descriptions are extremely terse (10-30 chars), parameter descriptions are minimal, and there is no documented output schema. The tools are intentionally fake ('performs no scan', 'fictitious record'), which limits practical utility. Naming follows verb_noun convention (create_scan, read_scan, complete_scan), but descriptions lack the depth needed for LLM tool selection in production. No error handling guidance, no output structure documentation, and no parameter validation beyond minLength constraints.

• Expand tool descriptions to 100-200 chars. Example: 'Creates an in-memory scan record for a target host. Returns a handle that can be used with read_scan and complete_scan. Note: this is a lab simulation and performs no actual security scan.'
• Document output schema for each tool. Example for read_scan: 'Returns {handle: string, target: string, note: string, status: "pending"|"completed", created_at: ISO8601}'
• Add parameter descriptions that explain format and constraints. Example for handle: 'Unique identifier returned by create_scan. Format: alphanumeric string, 1-128 chars.'
• Add error handling guidance. Example: 'If handle not found, returns 404 with message "Scan record not found. Call create_scan to create a new record."'
• Consider adding a list_scans tool to help agents discover valid handles without guessing.