The Agent Skills Directory

Remote Script Execution: The skill provides a command to download and execute a shell script from a remote GitHub repository (curl -fsSL ... | sh). This pattern executes code directly from an external source without a manual review step.
Indirect Prompt Injection Surface: The skill interacts with terminal output using screenshot and wait. If a TUI application displays untrusted data, that content enters the agent's context and could potentially affect its decision-making.
Ingestion points: Terminal content captured via agent-tui screenshot and agent-tui wait as described in SKILL.md.
Capability inventory: The skill can execute commands (run), provide keyboard input (type, press), and manage sessions (kill).
Boundary markers: No specific markers are used to differentiate terminal data from the agent's instructions.
Sanitization: The instructions do not specify any validation or filtering of the captured terminal text.
Background Service Management: The skill includes a workaround for macOS that uses tmux to run background daemons. This involves managing persistent sessions and temporary files to maintain the tool's connectivity.

agent-tui