GhostCheck exposes a single tool 'agencyauditor' with minimal definition quality. The tool name lacks a clear action verb (should be 'scan_files' or 'audit_agency'). Description is generic ('Scanner plugin for AgencyAuditor') and under 50 chars, providing no context for when/why an LLM should call it. Input schema is present but sparse: 'files' array and 'config' object lack detailed descriptions, type constraints, and usage guidance. No output schema is documented. No error handling guidance. The tool appears to be a thin wrapper around a CLI command with no LLM-specific optimization.

• Rename tool to 'scan_files' or 'audit_agency_code' to start with a clear action verb. This helps LLMs parse intent from the name alone.
• Expand description to 100-150 chars: 'Scans source files for security vulnerabilities using AgencyAuditor. Accepts file paths and optional config. Returns list of findings with severity, location, and remediation hints. Use this to audit code before deployment.'
• Replace 'config' object with explicit parameters: 'severity_threshold' (enum: low|medium|high|critical), 'check_types' (array of strings), 'timeout_seconds' (integer, 1-300). Add descriptions for each.
• Document output schema: 'Returns object with fields: vulnerabilities (array of {file, line, severity, message, remediation}), summary (string), scan_duration_ms (integer), total_issues (integer).'
• Add error handling: 'If files not found, returns error with list of invalid paths. If timeout exceeded, returns partial results with warning. If permission denied, suggests running with elevated privileges.'