Tmux

Security checks across static analysis, malware telemetry, and agentic risk

Overview

The tmux skill mostly matches its purpose, but it has an inconsistent socket boundary and promotes unattended coding-agent runs, so it deserves review before installation.

Install only if you are comfortable with the agent controlling tmux sessions. Use a private tmux socket, avoid unattended --yolo/full-auto coding-agent runs unless you explicitly approve them, review generated changes, and kill tmux sessions when finished.

Static analysis

No static analysis findings were reported for this release.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Risk analysis

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

#
ASI06: Memory and Context Poisoning
Medium
What this means

Private command output, tokens, or other sensitive text from a tmux pane could be brought into the agent transcript if the helper targets the wrong or ambient tmux session.

Why it was flagged

The helper captures pane history without accepting the isolated socket path used elsewhere in SKILL.md, and prints the captured pane text on timeout. This can expose terminal output from an unintended tmux context.

Skill content
tmux capture-pane -p -J -t "$target" -S "-${lines}" ... echo "Last ${lines} lines from $target:" >&2; printf '%s\n' "$pane_text" >&2
Recommendation

Add explicit -S/--socket-path support to the helper, require the private socket in examples, and avoid printing full pane history by default on timeout.

#
ASI02: Tool Misuse and Exploitation
Medium
What this means

Spawned coding agents may change files or run commands without step-by-step user confirmation.

Why it was flagged

The instructions explicitly promote running multiple coding agents through tmux in no-confirm or full-auto modes, without stating approval, diff review, rollback, or sandbox boundaries.

Skill content
tmux -S "$SOCKET" send-keys -t agent-1 "cd /tmp/project1 && codex --yolo 'Fix bug X'" Enter ... Codex needs `--yolo` or `--full-auto` for non-interactive fixes
Recommendation

Use --yolo/--full-auto only after explicit user approval, in disposable worktrees or sandboxes, and require reviewing diffs and command output before accepting changes.

#
ASI10: Rogue Agents
Low
What this means

Processes started in tmux may continue consuming resources or making changes until explicitly stopped.

Why it was flagged

The skill creates detached tmux sessions that can keep running after the immediate task, although cleanup commands are documented.

Skill content
tmux -S "$SOCKET" new -d -s "$SESSION" -n shell ... Kill a session: `tmux -S "$SOCKET" kill-session -t "$SESSION"`
Recommendation

Monitor active sessions and use the documented kill-session or kill-server cleanup commands when the task is complete.