The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes standard Node.js development commands (npm run build, npm test) to verify code integrity after implementing improvements. These operations are restricted to the local environment and are typical for development tools.
[DATA_EXFILTRATION]: The skill reads project source files (JavaScript, HTML, JSON) to perform its diagnostic audit. There are no network operations, external URLs, or indicators that any project data is transmitted outside the local environment.
[PROMPT_INJECTION]: While the skill processes external data by reading codebase files, it uses a structured multi-step workflow (Audit → Score → Plan → User Confirmation → Implement) that limits the impact of potential indirect prompt injection from malicious comments or content within the analyzed files.

improve-game