feat: auto experimentation table

[nnasko]

Summary

Hands-off Experimentation Table automation: walk to the table, play the add-ons then Superpairs, claim rewards, renew with bits + XP up to a daily cap, and resume farming.

Solvers (tick-driven, gated on a master toggle, driven off actual board state rather than phase-name matching):

• Superpairs — reveal cards, remember each slot, click a pair the moment both halves are known.
• Chronomatron — record each note as it lights up, replay the sequence (one note longer each round). No assumptions about repeats or ordering.
• Ultrasequencer — memorise the numbered panes, click them in order.

Session orchestrator (ExperimentationManager, new EXPERIMENTING macro state):

• Precise centred walk to a saved Table Spot, then open the table.
• Stakes-tier selection that skips enchanting-locked and unaffordable tiers, with practice-mode support (right-click, free, for iterating).
• XP-cost read from tier lore vs the player's level, with an optional buy + splash of a Titanic Experience Bottle when short.
• "Experiment Over" reward claiming, and renewal bounded by the server's own daily counter.
• Reward target / stop point read from the tier lore ("Series of 7" etc).

Wiring: config entries + a ServiceLoader UI page (Table Spot with in-world highlight, practice mode, click/note delays, renewal cap), MacroState state, tick-handler + chat hooks, and a design spec under docs/.

Testing

Verified in-game via practice mode with debug logging: navigation, tier selection, Superpairs and Ultrasequencer solving, reward claiming and renewal. Chronomatron solving was the last piece under active tuning against the board's note timeline.

Notes

Client-side GUI automation, so the container-menu details are matched from live board dumps; debug logging behind the existing Show Debug toggle narrates every decision for further tuning.