tmux-native-coding-agent

Pass

Audited by Gen Agent Trust Hub on Jun 13, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [COMMAND_EXECUTION]: The skill documents standard usage of tmux CLI tools (list-sessions, list-panes, capture-pane) to manage and inspect the terminal environment. These are legitimate administrative commands used for the stated purpose of coordinating developer workflows.
  • [DATA_EXFILTRATION]: The skill identifies an inherent risk in terminal-native workflows where tmux capture-pane might inadvertently read sensitive data such as API keys, tokens, or private logs present in the terminal buffer. It provides explicit safety boundaries and rules, instructing the agent to minimize scope, avoid indiscriminate capture, and seek user permission if secrets might be present.
  • [SAFE]: The skill is instructional and architectural in nature. It does not contain any executable code, remote dependencies, or hidden logic. Its primary function is to define a mental model for persistent AI pair-programming sessions.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 13, 2026, 01:59 AM
Security Audit — agent-trust-hub — tmux-native-coding-agent