The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides capabilities for running arbitrary shell commands via the terminal_exec tool, as well as executing commands inside containers and orchestration environments through docker_exec and k8s_exec.
[DATA_EXFILTRATION]: The skill includes tools to broadcast data across various platforms, including Slack (slack_send_message), WhatsApp (whatsapp_message), and Telegram (telegram_message). When combined with the skill's access to the filesystem and codebase, this creates a capability for exfiltrating sensitive local data to external services.
[PROMPT_INJECTION]: The skill processes data from various external and potentially untrusted sources which could be used for indirect prompt injection attacks.
Ingestion points: Data is retrieved from Shodan searches, CVE database details, GitHub repository content, and logs from Docker/Kubernetes environments.
Boundary markers: The instructions do not define any specific delimiters or boundary markers to separate retrieved data from agent instructions.
Capability inventory: The agent has high-privilege capabilities including filesystem access, shell execution, and the ability to send messages to multiple external network endpoints.
Sanitization: There is no evidence of input validation or sanitization for the data retrieved from external APIs before it is processed by the agent.

10x-developer