The Agent Skills Directory

[SAFE]: The skill provides legitimate technical implementation for game simulation logic. The code follows standard TypeScript patterns for state management and fixed-timestep loops, utilizing built-in functions like performance.now, TextEncoder, and structuredClone for deep object copying.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external data sources, specifically server state updates and player input sequences.
Ingestion points: The handleServerUpdate function in boilerplate/game-loop-client.ts and the reconcile method in boilerplate/state-sync.ts are the primary entry points for external data.
Boundary markers: None specific to prompt injection are present as the system operates on structured data (JSON/Binary) rather than natural language instructions.
Capability inventory: The provided scripts contain no high-risk capabilities; there are no network request functions, filesystem access commands, shell execution calls, or dynamic code evaluation (eval/exec).
Sanitization: The skill employs structuredClone to ensure state isolation and uses standard JSON.parse within its delta decoding logic.

engineering-game-state-sync