tamework-interaction-configurator

name	tamework-interaction-configurator
description	Use when editing Tamework `TwInteractionConfig` prompts, sensors, actions, cooldowns, contextual interactions, config-defined command modes, command cycles, or interaction-triggered role behavior. Does not own command-item hotswaps, panels, target authority, or HUD runtime.

Tamework Interaction Configurator

Keep interaction prompts and interaction execution perfectly aligned.

Workflow

1. Read references/interaction-checklist.md.
2. Read references/prompt-state-matrix.md.
3. Use $hytale-asset-tools and lock the exact project profile. Inspect the full sensor -> prompt -> action -> state change -> cooldown/reset path with declared/effective provenance, references, findings, and actionable advisories.
4. Use $hytale-asset-inheritance-contract for every TwInteractionConfig parent/fallback change. Do not flatten inherited values into children.
5. Use author options for role action/sensor/filter/state shapes and same-profile reviewed examples for orientation. Never author from a static type list.
6. Use hytale-workshop-mcp when command or interaction behavior depends on vanilla Hytale prompt, item interaction, sensor, action, target, cooldown, or role semantics not established by repo/profile evidence.
7. Build the config and role wiring as one read-only candidate. Run author validate --scope affected before materialization.
8. Generate verification for prompt visibility, executable action, denied conditions, mode cycle, cooldown, reset, target loss, and ownership where applicable. Use logs and live behavior only as additional runtime evidence.

Route Runtime Work

• Use $tamework-command-runtime when the change touches command-item hotswaps, radial controls, linked or bonded panels, target authority, HUD state, or command cleanup.
• Use $tamework-config-authoring for shared config-family schema, codec, inheritance, override, editor, cache, or reload behavior.
• Use $tamework-runtime-safety for ECS, tick, thread, async, or cadence work.

Output Contract

Return:

• Interaction files changed.
• Prompt/action alignment checks.
• Mode mapping table for touched interactions.
• Project profile, knowledge hash, snapshot/candidate identity, validation outcome, and verification gaps.

Guardrails

1. No prompt without executable action path.
2. No action path without gating sensor conditions.
3. Keep interaction type IDs stable unless explicitly breaking.
4. Distinguish vanilla interaction behavior from Tamework's optimized TwInteractionConfig path in reports.
5. Do not invent command modes, prompts, states, cooldown values, component interfaces, or inherited overrides.
6. Do not treat a visible prompt as proof that execution, reset, or ownership paths work.
7. Keep review-only fallback/inherited-default advisories distinct from deterministic wiring blockers.
8. Verify that the cooldown producer and the sensor/transition used for reset read the same alarm/state store. Similar names do not prove a shared lifecycle channel.