The Agent Skills Directory

[SAFE]: No malicious patterns or security vulnerabilities were detected across any of the provided files.
[PROMPT_INJECTION]: The skill's instructions are strictly limited to character development and empathetic response workflows. There are no attempts to bypass safety filters or override agent instructions.
[DATA_EXFILTRATION]: The skill does not contain any network-facing code or tools. It does not access local files, environment variables, or credentials.
[REMOTE_CODE_EXECUTION]: No external package dependencies or remote script downloads were identified. The skill is entirely declarative and text-based.
[COMMAND_EXECUTION]: The skill does not invoke shell commands or system processes. It lacks any capability to interact with the underlying operating system.

dazai-perspective