The Agent Skills Directory

[PROMPT_INJECTION] (SAFE): The instructions focus on organizational psychology and professional networking. No patterns attempting to bypass safety filters or override core agent behavior were detected.
[DATA_EXFILTRATION] (SAFE): No network operations, credential usage, or sensitive file path access (e.g., SSH keys, env files) are present in the skill.
[REMOTE_CODE_EXECUTION] (SAFE): There are no commands to download or execute external scripts, and no package managers (npm, pip) are invoked.
[COMMAND_EXECUTION] (SAFE): The skill does not contain any executable scripts, shell commands, or subprocess calls. It is entirely composed of markdown templates and guidance.
[NO_CODE] (SAFE): The skill is a pure documentation/template skill. Without executable logic or system interaction capabilities, it lacks the mechanisms required for traditional cyberattacks.

influence-craft