The Agent Skills Directory

[SAFE]: No malicious code, obfuscation, or unauthorized data access commands were detected in the skill files. The instructions are purely focused on game design logic and structured verification processes.
[INDIRECT_PROMPT_INJECTION]: The skill defines a process for ingesting external project data and user briefs. While this creates an input surface, the logic is constrained by a structured locking mechanism (Candidate Matrix, Test Protocol, Evidence Ledger, Identification Outcome) that serves as a boundary between data ingestion and implementation, effectively mitigating risks by requiring internal consistency checks before any state transitions.

worldview-game-evidence-based-entity-identification