The Agent Skills Directory

[SAFE]: The skill's primary function is information retrieval and synthesis from public web sources. It includes strong instructions for agents to remain in a passive extraction mode ('Return verbatim quotes') which effectively mitigates the risk of following malicious instructions found in public posts.
[DATA_EXFILTRATION]: While the skill collects public usernames and social media handles for the 'Named individuals to DM' section, it explicitly restricts its behavior to producing a static list for the user and strictly forbids automated contact or outreach. It does not access private files, environment variables, or sensitive credentials.
[REMOTE_CODE_EXECUTION]: No remote code execution, package installation, or shell command patterns were identified. The parallel agents described in the workflow represent standard LLM orchestration for task decomposition and do not involve the execution of untrusted scripts or binary files.

validate-icp