The Agent Skills Directory

[COMMAND_EXECUTION] (SAFE): The skill documents common Bash commands including process management (kill), file operations (rm), and text processing (sed, awk). These are standard utilities provided as educational patterns and do not contain instructions for malicious system interference.
[DATA_EXFILTRATION] (SAFE): The documentation includes curl for network requests and env for viewing environment variables. These are used in generic examples (e.g., https://example.com) and do not instruct the agent to access or transmit sensitive files like SSH keys or cloud credentials.
[REMOTE_CODE_EXECUTION] (SAFE): Although curl is mentioned, there are no patterns involving piping remote content directly into a shell (e.g., curl | bash), which is a common attack vector.

bash-linux