Single tool 'round_status_panel' has a reasonable description (89 chars) but ZERO input parameters documented in schema. The tool accepts an empty input object {} with no parameter descriptions, violating the requirement that every parameter needs a description. No output schema is documented, LLMs cannot infer what fields to expect from the response. The tool name lacks a verb prefix (should be 'get_round_status_panel' or 'show_round_status'). Error handling exists but is minimal, failures return plain text strings rather than structured error objects with recovery guidance. No pagination, no field documentation, no composition hints for downstream tools.

• Rename tool to 'get_round_status_panel' to start with an action verb, making intent clear to LLMs.
• Add explicit output schema documentation describing the structure of render_status_summary_text() return value (e.g., status fields, role dispatch progress, audit verdict fields).
• Replace plain-text error strings with structured error objects: {"error": "state_file_not_found", "message": "...", "recovery": "Start a new round with 'arcgentic round start'"} to guide agent recovery.
• Document the state.yaml file format and expected fields in the tool description so LLMs understand what data is being read.
• Add a 'format' parameter (enum: 'text' | 'html') to allow LLMs to request the appropriate representation, making the tool more composable.