Skills tmux
📦
tmux
Content revision r1 Medium Risk ⚙️ External commands📁 Filesystem access🔑 Env variables
Manage Persistent tmux Sessions
Long-running terminal tasks are hard to monitor across agent turns. This skill creates, lists, captures, controls, and cleans tmux sessions for persistent command execution.
Supports: Claude Codex Code(CC)
Install with my Agent
Copy this request to your Agent. It includes the canonical Skill page and manifest.
Agent request
Review the Skillstore skill "tmux" from https://skillstore.io/skills/dwsy-tmux.md and its manifest at https://skillstore.io/api/skills/dwsy-tmux/manifest. Verify the artifact. You may proceed after verification, subject to the environment's own policy.Your Agent should still show its plan and request any confirmation required by the security policy.
Agent-readable resources
Use these links when an AI agent, crawler, or script needs clean context instead of reading the full page.
Test it
Using "tmux". Start a background build and show how I can monitor it.
Expected outcome:
The skill creates a named tmux task session and returns the session identifier, attach command, and capture command for later monitoring.
Using "tmux". Check the latest output from my dev server session.
Expected outcome:
The skill captures recent pane text and summarizes whether the service appears ready, still starting, or failed.
Using "tmux". Clean old tmux sessions.
Expected outcome:
- Lists managed sessions with age and status.
- Removes sessions older than the requested threshold.
- Reports how many sessions were cleaned.
Security Audit
Medium Riskv9 • 7/9/2026 Open versioned report
Most static findings are false positives from Markdown code fences, TypeScript template strings, documented installation paths, and benign shell redirection. The real risk is intentional local command execution through tmux session creation and key sending, plus temporary socket and session metadata storage.
9
Files scanned
1,114
Lines analyzed
23
Review items
0
False positives ignored
Confirmed security concerns (1)
Medium
Arbitrary Local Command Execution by Design
The skill is designed to create tmux sessions from caller-provided commands and send caller-provided keys into live panes. This is useful for agent workflows but exposes the same risk as allowing an agent to run local shell commands.
Both implementation and documentation show command creation and keystroke sending into tmux sessions. This is intentional functionality, not hidden malware, but it is a real execution capability.
Capability review items (23)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
Medium
Ruby/shell backtick execution
await $`tmux -S ${socket} new -d -s ${id} -n shell`.quiet();
This line runs tmux or shell helper logic as part of the skill. The behavior is intentional, but it is a real local command-execution capability that should be disclosed.
Medium
Ruby/shell backtick execution
await $`tmux -S ${socket} send-keys -t ${target} -- ${command} Enter`.quiet();
This line runs tmux or shell helper logic as part of the skill. The behavior is intentional, but it is a real local command-execution capability that should be disclosed.
Medium
Ruby/shell backtick execution
await $`tmux -S ${session.socket} kill-session -t ${id}`.quiet();
This line runs tmux or shell helper logic as part of the skill. The behavior is intentional, but it is a real local command-execution capability that should be disclosed.
Medium
Ruby/shell backtick execution
await $`tmux -S ${socket} send-keys -t ${target} ${args}`.quiet();
This line runs tmux or shell helper logic as part of the skill. The behavior is intentional, but it is a real local command-execution capability that should be disclosed.
Medium
Ruby/shell backtick execution
const result = await $`tmux -S ${socket} capture-pane -p -J -t ${target} -S -${lines}`.quiet();
This line runs tmux or shell helper logic as part of the skill. The behavior is intentional, but it is a real local command-execution capability that should be disclosed.
Medium
Ruby/shell backtick execution
const result = await $`tmux -S ${socket} display-message -p '#{window_activity_flag}' -t ${target}`.
This line runs tmux or shell helper logic as part of the skill. The behavior is intentional, but it is a real local command-execution capability that should be disclosed.
Medium
Ruby/shell backtick execution
await $`tmux -S ${session.socket} kill-session -t ${id}`.quiet();
This line runs tmux or shell helper logic as part of the skill. The behavior is intentional, but it is a real local command-execution capability that should be disclosed.
Medium
Ruby/shell backtick execution
const result = await $`tmux -S ${socket} list-sessions -F '#{session_name}'`.quiet();
This line runs tmux or shell helper logic as part of the skill. The behavior is intentional, but it is a real local command-execution capability that should be disclosed.
Medium
Shell command substitution
if ! sessions="$("${tmux_cmd[@]}" list-sessions -F '#{session_name}\t#{session_attached}\t#{session_
This line runs tmux or shell helper logic as part of the skill. The behavior is intentional, but it is a real local command-execution capability that should be disclosed.
Medium
Shell command substitution
sessions="$(printf '%s\n' "$sessions" | grep -i -- "$query" || true)"
This line runs tmux or shell helper logic as part of the skill. The behavior is intentional, but it is a real local command-execution capability that should be disclosed.
Medium
Shell command substitution
attached_label=$([[ "$attached" == "1" ]] && echo "attached" || echo "detached")
This line runs tmux or shell helper logic as part of the skill. The behavior is intentional, but it is a real local command-execution capability that should be disclosed.
Medium
Shell command substitution
start_epoch=$(date +%s)
This line runs tmux or shell helper logic as part of the skill. The behavior is intentional, but it is a real local command-execution capability that should be disclosed.
Medium
Shell command substitution
deadline=$((start_epoch + timeout))
This line runs tmux or shell helper logic as part of the skill. The behavior is intentional, but it is a real local command-execution capability that should be disclosed.
Medium
Shell command substitution
pane_text="$(tmux capture-pane -p -J -t "$target" -S "-${lines}" 2>/dev/null || true)"
This line runs tmux or shell helper logic as part of the skill. The behavior is intentional, but it is a real local command-execution capability that should be disclosed.
Medium
Shell command substitution
now=$(date +%s)
This line runs tmux or shell helper logic as part of the skill. The behavior is intentional, but it is a real local command-execution capability that should be disclosed.
Medium
Temp directory access
- **Default Socket**: `/tmp/pi-tmux-sockets/pi.sock`
The documentation references tmux sockets or managed state under /tmp. Temporary filesystem usage is expected, but it is still a real local filesystem dependency.
Medium
Temp directory access
# Output: tmux -S /tmp/pi-tmux-sockets/pi.sock attach -t pi-task-python-20250107-123456
The documentation references tmux sockets or managed state under /tmp. Temporary filesystem usage is expected, but it is still a real local filesystem dependency.
Medium
Temp directory access
"socket": "/tmp/pi-tmux-sockets/pi.sock",
The documentation references tmux sockets or managed state under /tmp. Temporary filesystem usage is expected, but it is still a real local filesystem dependency.
Medium
Temp directory access
tmux -S /tmp/pi-tmux-sockets/pi.sock attach -t pi-service-dev-server-20250107-123456
The documentation references tmux sockets or managed state under /tmp. Temporary filesystem usage is expected, but it is still a real local filesystem dependency.
Medium
Temp directory access
tmux -S /tmp/pi-tmux-sockets/pi.sock attach -t pi-task-compile-20250107-123456
The documentation references tmux sockets or managed state under /tmp. Temporary filesystem usage is expected, but it is still a real local filesystem dependency.
Medium
Temp directory access
tmux -S /tmp/pi-tmux-sockets/pi.sock capture-pane -p -J -t pi-task-compile-20250107-123456:0.0 -S -2
The documentation references tmux sockets or managed state under /tmp. Temporary filesystem usage is expected, but it is still a real local filesystem dependency.
Medium
Temp directory access
tmux -S /tmp/pi-tmux-sockets/pi.sock attach -t <session-id>
The documentation references tmux sockets or managed state under /tmp. Temporary filesystem usage is expected, but it is still a real local filesystem dependency.
Medium
Temp directory access
ls -la /tmp/pi-tmux-sockets/
The documentation references tmux sockets or managed state under /tmp. Temporary filesystem usage is expected, but it is still a real local filesystem dependency.
Risk Factors
⚙️ External commands (50)
examples/long-task.ts:15 examples/long-task.ts:16 examples/long-task.ts:17 examples/long-task.ts:18 examples/long-task.ts:19 examples/python-repl.ts:11 examples/python-repl.ts:12 examples/python-repl.ts:13 examples/start-service.ts:15 examples/start-service.ts:16 examples/start-service.ts:17 examples/start-service.ts:18 examples/start-service.ts:19 examples/start-service.ts:33 lib.ts:41 lib.ts:45 lib.ts:89 lib.ts:90 lib.ts:120 lib.ts:122 lib.ts:130 lib.ts:135 lib.ts:151 lib.ts:152 lib.ts:159 lib.ts:185 lib.ts:201 lib.ts:233 lib.ts:234 lib.ts:235 lib.ts:236 lib.ts:237 lib.ts:252 lib.ts:253 lib.ts:267 lib.ts:271 lib.ts:272 lib.ts:273 lib.ts:274 lib.ts:275 lib.ts:276 lib.ts:288 lib.ts:293 lib.ts:305 lib.ts:321 lib.ts:329 lib.ts:341 lib.ts:344 lib.ts:345 lib.ts:346
📁 Filesystem access (50)
examples/long-task.ts:2 examples/long-task.ts:7 examples/python-repl.ts:2 examples/python-repl.ts:7 examples/start-service.ts:2 examples/start-service.ts:7 examples/start-service.ts:33 examples/start-service.ts:33 scripts/find-sessions.sh:46 scripts/find-sessions.sh:55 scripts/wait-for-text.sh:57 scripts/wait-for-text.sh:68 scripts/wait-for-text.sh:70 SKILL.md:15 SKILL.md:18 SKILL.md:21 SKILL.md:24 SKILL.md:53 SKILL.md:60 SKILL.md:65 SKILL.md:72 SKILL.md:73 SKILL.md:78 SKILL.md:85 SKILL.md:90 SKILL.md:97 SKILL.md:102 SKILL.md:109 SKILL.md:114 SKILL.md:121 SKILL.md:126 SKILL.md:133 SKILL.md:134 SKILL.md:139 SKILL.md:146 SKILL.md:152 SKILL.md:191 SKILL.md:194 SKILL.md:197 SKILL.md:200 SKILL.md:208 SKILL.md:211 SKILL.md:212 SKILL.md:220 SKILL.md:226 SKILL.md:234 SKILL.md:237 SKILL.md:238 SKILL.md:256 SKILL.md:299
Audited by: claude View Audit History →
Share & cite this report
Share the versioned assessment report, neutral badge, embed card, and citations. Skillstore reports evidence without deciding whether this Skill is safe.
Copy report link
https://skillstore.io/skills/dwsy-tmux/audits/9?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_reportMarkdown badge
[](https://skillstore.io/skills/dwsy-tmux?utm_source=security_passport_badge)HTML badge
<a href="https://skillstore.io/skills/dwsy-tmux?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/dwsy-tmux/security.svg" alt="Skillstore security assessment" loading="lazy"></a>Embed card
<iframe src="https://skillstore.io/embed/skills/dwsy-tmux.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>Academic citations (APA · BibTeX · CFF)
APA citation
Dwsy. (2026). tmux security audit report (audit version 9) [Author version unspecified]. Skillstore. https://skillstore.io/skills/dwsy-tmux/audits/9BibTeX citation
@techreport{dwsy-dwsy-tmux-2026,
author = {Dwsy},
title = {tmux security audit report (audit version 9)},
institution = {Skillstore},
year = {2026},
number = {9},
url = {https://skillstore.io/skills/dwsy-tmux/audits/9},
note = {Author version unspecified}
}CITATION.cff
cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "tmux security audit report (audit version 9)"
version: "unspecified"
type: report
authors:
- name: "Dwsy"
date-released: "2026-07-09"
url: "https://skillstore.io/skills/dwsy-tmux/audits/9"
identifiers:
- type: other
value: "skillstore:dwsy-tmux:audit:9"
description: "Skillstore immutable audit report identifier"
Compare variants
2 installable variantsEach author remains a separate installable skill. The recommended variant is ranked by Skillstore evidence.
Why this variant is first
Highest Skillstore Score
Skillstore Score
Why this score Evidence Confidence: High 45
Architecture
85
Maintainability
87
Content
71
Community
83
Spec Compliance
What You Can Build
Run long builds outside the main agent turn
Start a build or test command in tmux, then capture output later without losing process state.
Manage local development services
Launch a dev server in a named tmux session and inspect logs while other work continues.
Drive interactive terminal programs
Start a REPL or CLI, send input, and capture responses from the same persistent pane.
Try These Prompts
Start a simple task
Use the tmux skill to start a task named build that runs my project build command. Show me how to capture the latest output.
Monitor a dev server
Create a tmux service session for my development server, then check whether the server started successfully.
Control an interactive REPL
Start a Python REPL in tmux, send a simple expression, and capture the response from the pane.
Clean stale sessions safely
List all tmux skill sessions, identify inactive sessions older than two days, and clean only those sessions.
Best Practices
- Review commands before creating sessions because they run in your local shell context.
- Use descriptive session names so cleanup and monitoring are clear.
- Capture output before killing a session if logs may be needed later.
Avoid
- Do not run untrusted commands or paste secrets into managed sessions.
- Do not leave long-lived services running without periodic cleanup.
- Do not assume tmux sessions isolate commands from your local filesystem.
Frequently Asked Questions
What does this skill require?
It requires tmux, Bun, and a local environment where shell commands may run.
Can it run interactive commands?
Yes. It can send keys to a tmux pane and capture the pane output later.
Does it sandbox commands?
No. Commands run with the same local permissions as the user or agent process.
Where are sessions tracked?
The skill stores session metadata and sockets in local temporary or configured filesystem paths.
Can Claude Code use it for dev servers?
Yes. It can start a server in tmux and inspect output while the agent continues other work.
Is it safe for untrusted input?
No. Treat commands and keystrokes as executable input and review them before use.