The Agent Skills Directory

[DYNAMIC_CONTEXT_INJECTION]: The skill uses the !command syntax in SKILL.md to load shared protocol files and configuration settings (e.g., cat skills/_shared/protocols/ux-protocol.md). These commands facilitate modular context loading at skill initialization and do not involve sensitive file access or network exfiltration.- [COMMAND_EXECUTION]: The LITE.md file contains instructions for the agent to execute shell commands such as cat package.json and find assets/ to ground itself in the local project environment. These are standard discovery steps for a development tool.- [INDIRECT_PROMPT_INJECTION]: The skill identifies an ingestion surface for indirect prompt injection. 1. Ingestion points: PIPELINE_CONTEXT.visual_basis and Visual Evidence Cards referenced in SKILL.md. 2. Boundary markers: No specific delimiters or "ignore instructions" warnings are present in the guidelines. 3. Capability inventory: The skill can generate Phaser JavaScript/TypeScript code and execute local discovery commands. 4. Sanitization: No sanitization or validation of the input context is described. The severity is low as the primary focus is asset generation.

game-asset-vfx