Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/HLH2023/terminal-use-mcpnpx agentmods add skills/hlh2023/terminal-use-mcp/tui-claude-codeWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/hlh2023/terminal-use-mcp/tui-claude-code)<a href="https://agentmods.dev/skills/hlh2023/terminal-use-mcp/tui-claude-code"><img src="https://agentmods.dev/badge/skills/hlh2023/terminal-use-mcp/tui-claude-code/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/hlh2023/terminal-use-mcp/tui-claude-code"><img src="https://agentmods.dev/badge/skills/hlh2023/terminal-use-mcp/tui-claude-code.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00031 | $0.01875 |
| Opus 5 | $0.00015 | $0.00937 |
| Sonnet 5 | $0.00006 | $0.00375 |
| Haiku 4.5 | $0.00003 | $0.00187 |
Grade A, and why
tui-claude-code scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 197 lines — stays where its author put it; the contents beside it link to each section on GitHub.
tui-claude-code: Control Claude Code TUI
Reference: Claude Code v2.1.179 — Keybindings verified against this version. This skill is community-maintained; it is NOT updated in lockstep with Claude Code releases. If Claude Code updates break keybindings, update this skill yourself or submit a PR to GitHub.
This skill is optional. Only install if you need to control this agent's TUI via terminal-use-mcp.
This skill is useful when one AI agent needs to remotely control a Claude Code TUI session via terminal-use-mcp.
Operational skill for controlling the Claude Code TUI through terminal-use-mcp. Includes complete key mappings, slash commands, permission modes, and interaction flows.
When To Use
- You need to start and interactively control a Claude Code TUI through terminal-use-mcp.
- You need to switch permission modes or run slash commands inside Claude Code.
- You need to read Claude Code conversation content or search historical messages.
- You need advanced features such as the transcript viewer,
/btwside questions, or background tasks.
Core Operation Flow
Start And Readiness Check
1. terminal.start(command="claude", cwd="~/project")
2. terminal.wait_stable(idleMs=5000, timeoutMs=15000)
3. terminal.find("claude|Hi|What") # Confirm the UI is ready
Send A Message
terminal.type("your question")
terminal.press("enter")
terminal.wait_stable(idleMs=15000, timeoutMs=120000) # Wait for the model response
Interrupt And Exit
terminal.press("escape") # Interrupt the current response/tool call
terminal.press("ctrl+c") # Interrupt input; press a second time to exit
terminal.press("ctrl+d") # Exit
Global Keybindings
| Key | Function | Notes |
|---|---|---|
ctrl+c |
Interrupt / clear input | First press clears input, second press exits |
escape |
Interrupt current response/tool call | Stops active generation |
escape+escape |
Clear draft / open rewind | Double Esc opens rewind when input is empty |
ctrl+d |
Exit | Exit directly |
ctrl+l |
Redraw screen | Refresh terminal display |
ctrl+o |
Transcript viewer | View/browse the full conversation record |
ctrl+r |
History search | Search previous commands |
ctrl+b |
Background current task | Move the current task to the background |
ctrl+t |
Task list | View background tasks |
shift+tab |
Cycle permission mode | Cycles default -> acceptEdits -> plan |
alt+p / option+p |
Switch model | Open model picker |
alt+t / option+t |
Toggle extended thinking | Enable/disable deeper thinking |
alt+o / option+o |
Toggle fast mode | Fast mode |
ctrl+g / ctrl+x+ctrl+e |
External editor | Edit the message in an external editor |
ctrl+j |
Insert newline | Add a line break in the input box |
ctrl+v / cmd+v / alt+v |
Paste image | Paste an image from the clipboard |
up/down or ctrl+p/ctrl+n |
Cursor movement / history | Move up/down or browse message history |
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 9d ago First seen · 197 lines · 31 tokens per session scan A 82b6e16b3515
tui-claude-code is a skill published in the GitHub repository HLH2023/terminal-use-mcp (0 stars, last pushed 2mo ago), licensed MIT. It adds 31 tokens to every session and 1,875 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
data-analysis
Analyze, explore, clean, and visualize datasets with statistical rigor. Use when user asks to analyze data, find patterns, compute statistics, create visualizations, clean messy data, or explore a dataset. Trigger when user says things like "analyze this data", "what trends do you see", "find patterns in", "create a…
summarization
Summarize documents, articles, conversations, code, and technical content into concise, accurate summaries. Use when user asks to summarize, condense, create a TL;DR, write an executive summary, extract key points, or distill content. Trigger when user says things like "summarize this", "give me the key points"…
experiment_management
Set up and manage the experiment folder structure. This is Phase 0 — it runs before any analysis begins. All bookkeeping files are JSON (never markdown).
evaluate
Compare baseline and new implementation results. Produce the machine-readable final report result.json, update experiments.json, and append a row to comparison.json.
computer-use-linux
Linux desktop observation and control via native Pi tools or the computer-use-linux MCP server: accessibility trees, screenshots, window targeting, and input synthesis (click, type, scroll).
implement
Create a new Jupyter notebook implementing the method from the research paper, using the same data as the baseline. Record implementation details and measured metrics as a structured JSON entry.