The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes the claude CLI and system utilities like lsof via subprocess calls to perform evaluations and manage the local results viewer server. These actions are documented and align with the tool's primary purpose.
[SAFE]: External dependencies such as the anthropic client and the SheetJS library are well-known and reputable.
[SAFE]: The skill implements safety-conscious instructions, advising the agent to avoid creating malicious or misleading content. No evidence of data exfiltration or credential harvesting was found.

skill-creator