The Agent Skills Directory

[SAFE]: The skill consists entirely of instructional text and methodology derived from the book 'Contagious' by Jonah Berger. No malicious patterns, obfuscation, or attempts to bypass safety filters were found.
[NO_CODE]: There are no scripts, binaries, or configuration files that execute logic. The skill operates purely through prompt engineering to guide the AI's diagnostic reasoning.
[DATA_EXFILTRATION]: No network operations or external URLs were detected. The skill does not access local files or environment variables.
[PROMPT_INJECTION]: The instructions do not contain 'ignore previous instructions', 'system prompt extraction', or jailbreak-style commands. The logic follows a standard educational/consulting format.
[COMMAND_EXECUTION]: No shell commands or dynamic execution patterns were identified in the markdown or the test cases.

stepps-diagnostic