The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the mori CLI and tmux to manage terminal panes and execute commands. Key operations include creating new panes with mori pane new, reading pane history with mori pane read or tmux capture-pane, and sending input to processes using tmux send-keys.
[PROMPT_INJECTION]: The bridge design creates a surface for indirect prompt injection by enabling the agent to read and process output from other terminal panes which might contain untrusted content.
Ingestion points: Data enters the agent's context through mori pane read and tmux capture-pane operations documented in SKILL.md.
Boundary markers: There are no instructions for implementing boundary markers or system-level directives to ignore embedded instructions within the data read from other panes.
Capability inventory: The agent is granted capabilities to manage pane lifecycles and send arbitrary commands/input to any pane in the session via tmux send-keys.
Sanitization: The skill does not provide methods for sanitizing, validating, or escaping the content retrieved from panes before it is processed.

mori-agent-bridge