The Agent Skills Directory

[PROMPT_INJECTION]: No attempts to override agent behavior, bypass safety guidelines, or extract system prompts were detected. The instructions focus entirely on psychological analysis and user education.
[DATA_EXFILTRATION]: The skill does not contain any network operations, hardcoded credentials, or access to sensitive local file paths (e.g., .ssh, .aws, .env).
[OBFUSCATION]: No Base64, zero-width characters, homoglyphs, or encoded content were found. The text is clear and readable in both Chinese and English.
[REMOTE_CODE_EXECUTION]: There are no package installations (npm/pip) or patterns of downloading and executing remote scripts (curl/wget | bash).
[DYNAMIC_CONTEXT_INJECTION]: The skill does not use the shell execution syntax (!command) to run commands at load time.
[INDIRECT_PROMPT_INJECTION]: While the skill processes user-supplied scenarios for analysis, it lacks dangerous capabilities such as file writing, network access, or dynamic code execution that could be leveraged by malicious input. The risk surface is limited to text-based psychological interpretation.

contrast-principle