The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill provides instructions for the agent to install additional capabilities via the command npx skills add "" -y -g. This allows the installation and execution of code from a remote repository URL provided at runtime.
[COMMAND_EXECUTION]: The skill relies on the execution of multiple shell commands such as vs auth, vs llm, vs doctor, and app status to manage local environment configuration and verify operational readiness.
[DATA_EXPOSURE]: The skill manages sensitive credentials including VIKING_AK, VIKING_SK, and LLM API keys. It follows security best practices by explicitly instructing the agent to import these from environment variables or use interactive terminal logins to save them to a local secure store, strictly forbidding the agent from asking the user to provide secrets directly in the chat interface.
[INDIRECT_PROMPT_INJECTION]: The skill contains a multi-step chain instruction requiring the agent to consult another skill (vs-product-qa) to verify command syntax before execution. This introduces a surface where output from the external skill influences agent behavior, though the risk is low given the administrative nature of the task.

vs-shared