The Agent Skills Directory

[SAFE]: The skill consists of instructions and TypeScript code examples for refactoring purposes. No malicious patterns, obfuscation, or data exfiltration techniques were identified.
[PROMPT_INJECTION]: There are no instructions that attempt to bypass safety filters or override system constraints. The instructions are focused on code analysis and planning.
[COMMAND_EXECUTION]: The skill does not instruct the agent to execute shell commands or binary files. It explicitly states that the agent should only produce a plan and not execute it without human confirmation.
[DATA_EXFILTRATION]: The skill contains no network-reaching code. Examples of URLs and database connection strings in the evaluation files (e.g., https://api.example.com, postgres://main-db:5432/app) are standard placeholders used for testing and documentation.

nullables-refactor