egorfedorov AI Slot Game Developer
Use this skill to ship AI slot features that stay deterministic, performant, and safe under production constraints.
Workflow
- Define slot runtime contract.
- Specify AI feature scope per mode (
base,bonus,buy, etc.), update points, and max runtime latency. - Declare deterministic fallback behavior for each mode.
- Implement AI adapters and mode bindings.
- Keep AI provider/model calls behind adapter interfaces.
- Bind each AI system to explicit slot mode and spin-cycle stage.
- Keep payout-critical logic deterministic and provider-independent.
- Enforce runtime and safety budgets.
- Apply timeout, queue-depth, and retry limits.
- Ensure inference failure cannot block spin resolution.
- Add safe default outputs per mode.
- Validate integration spec.
- Validate modes/systems/models/runtime/safety/telemetry consistency.
- Validate fallback model references and mode references.
- Treat missing fallback or missing telemetry as blockers.
- Prepare release handoff.
- Deliver runtime map, mode budgets, fallback plan, and patch targets.
- Include exact verification commands and acceptance criteria.
Commands
python3 scripts/validate_ai_slot_runtime_spec.py \
--input <path/to/ai_slot_runtime_spec.json>
Treat non-zero exits as blocker findings.
Output Contract
Return:
Slot Runtime Map: modes, AI systems, models, and budgets.Validation Findings: pass/fail with exact mismatches.Patch Plan: modules/config files to update.Verification: command outputs and pass criteria.Residual Risks: unresolved runtime or safety concerns.
References
references/workflow.md: implementation-to-release process.references/slot-runtime-rules.md: slot-specific runtime and safety guardrails.references/signoff-template.md: release sign-off template.
Execution Rules
- Keep payout-critical path deterministic regardless of model outputs.
- Keep mode fallback behavior explicit and tested.
- Keep runtime budgets bounded and measurable.
- Flag unsafe fallback or missing telemetry as blockers.