The Agent Skills Directory

[SAFE]: The skill implements a robust integrity verification mechanism via scripts/check.py. This script is designed to run locally and validates the package structure, reference links, and documentation consistency without engaging in network activity or accessing sensitive host data.\n- [SAFE]: The instructions contain proactive security guidelines, specifically mandating that the agent must not follow instructions found within target project assets or metadata, and explicitly forbidding secret exposure or unauthorized external writes.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted project files and asset metadata, which serves as an ingestion point for potential indirect prompt injection. However, this risk is thoroughly mitigated by clear boundary instructions that define external content as data only and prohibit the execution of embedded commands (Ingestion points: Project source code, assets, metadata. Boundary markers: Explicit instructions in SKILL.md to ignore embedded commands. Capability inventory: No network or write tools granted to the skill. Sanitization: Instructions to treat all external inputs as inert data).

threejs-to-easeljs