Three tools with basic schemas but significant quality gaps. RaidShadowLegendsControl and RaidShadowLegendsSetGoal have enum-constrained parameters and reasonable descriptions (60-80 chars), but RaidShadowLegendsStatus has an empty input schema and minimal description (40 chars). Parameter descriptions are sparse or missing context. No output schemas documented. Error handling is absent, no guidance on retryability, user-fixable errors, or recovery paths. Tool names lack action verbs (Control/Status/SetGoal are weak; should be execute_raid_command, get_raid_status, configure_raid_goal). The 'parameters' field in RaidShadowLegendsControl is a generic record with no structure, forcing LLMs to guess valid keys. No mention of side effects, prerequisites, or when to use each tool.

• Rename tools with action verbs: execute_raid_command, get_raid_status, configure_raid_goal. Align with verb_noun baseline (90% of A+ tools start with action verb).
• Document output schemas for all tools. RaidShadowLegendsStatus should declare: {status: string, progress: number, battleCount: number, energyRemaining: number, lastUpdate: string, currentActivity?: string, error?: string}.
• Replace generic 'parameters' record in RaidShadowLegendsControl with explicit optional fields (e.g., duration?: number, intensity?: enum) or split into separate tools (start_raid, toggle_auto_raid, end_raid, lock_raid_screen).
• Add error handling guidance: 'If MQTT broker is unreachable, retry after 5 seconds. If command is invalid, check enum values. If Raid app is not running, start it first.'
• Expand RaidShadowLegendsStatus description from 40 chars to 100-150 chars explaining what status fields mean and when to call it (e.g., 'Call before each command to verify automation is ready').