The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze external, untrusted source code, which represents a potential surface for indirect prompt injection attacks.
Ingestion points: The workflow involves reading and analyzing arbitrary user-provided code files using tools like Read, Glob, and Grep (SKILL.md).
Boundary markers: The instructions do not define specific delimiters or "ignore instructions" wrappers for the code being reviewed, which could theoretically allow instructions within the source code to influence the agent's behavior.
Capability inventory: The skill is granted Write and Edit permissions (SKILL.md), meaning that if an agent were successfully manipulated by embedded instructions, it could potentially modify files in the environment.
Sanitization: There are no explicit sanitization or pre-processing steps defined for the code content before it is processed by the agent.
[EXTERNAL_DOWNLOADS]: The skill references external community-maintained Model Context Protocol (MCP) servers for engine-specific analysis (e.g., github.com/Coding-Solo/godot-mcp, github.com/CoplayDev/unity-mcp, and github.com/chongdashu/unreal-mcp). These references are provided as resources for extending the workflow's capabilities.

game-code-review