Single tool 'warmline_scan' has a basic schema with one parameter but critical gaps. Description is present (94 chars) but lacks actionable context for LLM selection. Parameter 'target' has type and description but no constraints (format, examples, or validation rules). No output schema documented, LLM cannot predict return structure. No error handling guidance. Tool does one thing (scan), but definition quality is below production baseline.

• Document output schema in tool description or as a separate schema field: {findings: [{id, sev, where}], score: number, tool: string}. This enables LLM to chain results to downstream tools.
• Add parameter constraints to 'target': specify it accepts absolute or relative directory paths, max length, and whether symlinks are followed. Example: 'Absolute or relative path to directory to scan (max 1024 chars, symlinks followed).'
• Expand tool description to 150-200 chars: explain that it scans for code markers (TODO, FIXME, XXX) and ranks findings by severity, when to call it (pre-commit, CI gates), and what the score represents.
• Add error handling guidance: 'If path is invalid or unreadable, returns empty findings. If YAML rulebook is malformed, returns error with line number. Call with a valid directory path.'
• Rename tool to 'scan_for_code_findings' or 'scan_directory_for_markers' to make the action and domain explicit to LLMs.