The Agent Skills Directory

[SAFE]: The skill instructions and reference materials do not contain any malicious patterns, unauthorized network requests, or sensitive data access.
[PROMPT_INJECTION]: The skill contains directives that modify the agent's standard behavior, such as "No questions. Do not use the question tool" and "No early exits." While these override default interaction patterns to facilitate an "autonomous mode," they are consistent with the skill's stated purpose and do not attempt to bypass safety guidelines or security filters.
[COMMAND_EXECUTION]: The documentation mentions standard project maintenance commands (e.g., pnpm test, pnpm build, pnpm lint) as examples of verification. These are expected behaviors within a developer tools context and do not constitute arbitrary or dangerous command execution.
[DATA_EXFILTRATION]: There is no evidence of attempts to harvest credentials or exfiltrate data. An example preview URL (https://preview-123.app) is provided in the documentation for illustrative purposes only.

engage