What problem does it solve? Running a fleet of Claude agents across multiple accounts makes it hard to know how much token quota remains, which pool is active, and when to rotate accounts before hitting a hard stop. This Skill turns scattered usage readings into a single estate table and one actionable verdict. ## Core Features & Use Cases - Estate-wide quota reporting: Reads live usage meters from every account pool and presents them in one table with runway, trust level, and reset times. - Trend logging: Appends a one-line entry per run to a trend log so quota burn can be tracked over days. - Rotation verdicts: Distinguishes normal burn from genuine Tier 2 rotation calls and re-pings open decisions as their projected date approaches. - Use Case: Three times a day, an automated check reads the active account's usage, records both weekly and session meters via fleet_budget_watch.py, and reports whether the current pool outlasts the next pool's reset or needs a rotation. ## Quick Start Ask the agent to run the token watch check and report the current estate status and rotation verdict.
