tmux

Pass

Audited by Gen Agent Trust Hub on Mar 2, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill allows for the remote control of tmux sessions, including sending literal keystrokes and commands via tmux send-keys. This is a primary feature intended for orchestrating interactive terminal sessions and complex CLI tools.
  • [PROMPT_INJECTION]: The skill facilitates indirect prompt injection by design, as it captures terminal output and uses it to drive agent logic (e.g., polling for specific text strings or monitoring process completion).
  • Ingestion points: Terminal output is ingested from tmux panes using tmux capture-pane in SKILL.md and scripts/wait-for-text.sh.
  • Boundary markers: There are no boundary markers or instructions to ignore embedded commands in the captured terminal output.
  • Capability inventory: The agent has the capability to execute any terminal command within the tmux session via send-keys.
  • Sanitization: The captured terminal output is not sanitized or escaped before being processed by regex matching or used in agent decision-making loops.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 2, 2026, 08:45 AM