The Agent Skills Directory

[COMMAND_EXECUTION]: The skill launches shell processes and SSH clients using subprocess.Popen. These executions use argument lists rather than raw shell strings, which is a best practice to prevent shell injection vulnerabilities.
[COMMAND_EXECUTION]: The script implements a background server model where a second instance of the script is launched via subprocess.Popen to manage the terminal session. It uses a local TCP socket bound to 127.0.0.1 for inter-process communication between the CLI and the background server.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection.
Ingestion points: External data enters the agent context via terminal output read in scripts/terminal_session.py (via ProcessTransport._reader and SerialTransport._reader).
Boundary markers: None are used when returning terminal output to the agent.
Capability inventory: The skill can execute local shell commands and SSH commands via subprocess.Popen in scripts/terminal_session.py.
Sanitization: No filtering or sanitization of terminal output is performed before it is passed to the agent.
[DATA_EXFILTRATION]: The skill manages session logs and state files within a local workspace directory (.embeddedskills/). No unauthorized data transmission to external domains was detected, and networking is limited to the local loopback interface for session control.

terminal