tmux

Pass

Audited by Gen Agent Trust Hub on Mar 1, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill facilitates arbitrary command execution in terminal sessions via the tmux send-keys command. This provides the agent with the ability to interact directly with the underlying shell and execute any command.
  • [DATA_EXFILTRATION]: Using tmux capture-pane, the skill can read and expose the entire history of a terminal session. This could lead to the exposure of sensitive data, such as credentials, session tokens, or private configuration details displayed in the terminal.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted data from terminal outputs.
  • Ingestion points: The scripts/wait-for-text.sh script and instructions in SKILL.md (via capture-pane) read raw text directly from tmux panes.
  • Boundary markers: Absent. There are no delimiters or instructions provided to the agent to ignore embedded commands or instructions within the captured terminal text.
  • Capability inventory: High. The skill can send arbitrary keystrokes to a shell (send-keys), create or kill sessions, and manage background processes.
  • Sanitization: Absent. Captured terminal output is processed and interpreted by the agent without any sanitization, escaping, or validation.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 1, 2026, 05:15 AM