dsh-ssh-tui

Usage

Key	Action
Enter	send the message; while the agent is running, steers it
Tab	complete the highlighted slash command
↑ / ↓	navigate slash-command suggestions (or input history)
Esc / Ctrl+C	cancel the running turn
Ctrl+D	exit
Ctrl+L	redraw
Ctrl+T	fold/unfold the input box (display-only; submission keeps the full text)
↑ / ↓	input history
y / n / Esc	answer an approval prompt
1..9 + Enter	answer an ask_user_question dialog

Type / to see slash-command suggestions — the panel merges the TUI's own commands with every command the harness registers (/goal, /plan, /compact, /permission, /feedback, ...). Tab completes, Enter runs. /help lists everything.

/model opens a two-step selector: pick a model from the current provider's catalog, then pick its reasoning effort (off / high / max when the provider exposes them). The change applies to the next request without changing the provider, updates the header/status line, and is remembered in agent-default-model for future launches.

/usage (alias /quota) works when the current provider is an OpenCode source and keeps the two billing models distinct:

• OpenCode Go queries the official quota endpoint and shows rolling 5-hour / weekly / monthly usage percentages, limit state, and reset times;
• OpenCode Zen is metered per API bill and has no fixed quota, so the TUI points to https://opencode.ai/zen for balance/billing and shows the session token usage it has recorded instead of inventing a quota.

The startup screen shows the official DeepSeek whale logo (rendered from the harness favicon) in the DeepSeek brand color, with the wordmark below it. The logo scales to the terminal width — a 52-column variant on wide terminals, down to a compact variant on narrow ones — so it never looks squeezed. A horizontal rule separates the workspace (transcript, reasoning, tool cards) from the input area.

Model reasoning blocks are collapsed by default: while thinking a compact ▸ 思考中 ⠹ · N 字 · Ns line with a spinner replaces the raw stream, and after the turn each block collapses to a ▸ 已思考 · N 行 summary without its content. The thinking block can be expanded live while streaming to watch the raw reasoning as it arrives. Assistant replies render in bold white with terminal markdown support: heading levels (H1 enlarged/underlined, H2 underlined, H3 colored), bold, italic, inline code, fenced code blocks, lists, quotes, and links all get ANSI styling while remaining width-wrapped for the terminal. Reasoning and tool cards are each expandable/collapsible independently — Ctrl+N / Ctrl+P move the selection highlight between them, Ctrl+R expands/collapses all blocks at once (individual blocks use ↑/↓ + Enter), and Esc drops the selection. With the input box empty, ↑/↓ move the selection and Enter toggles the selected block directly. Clicking a reasoning or tool header in the transcript also toggles it.

The transcript is scrollable: PgUp/PgDn or the mouse wheel move back through earlier reasoning blocks and tool calls, a ↑ 已回看 N 行 indicator shows the scroll position, and Esc (or sending a message) returns to the live bottom.

The terminal window title mirrors the session state while unfocused: an animated spinner plus 运行中 · 工具 N while working, ✓ 已完成 for a few seconds after completion, and 待命 when idle. A terminal bell rings on completion (DSH_TUI_NO_BELL=1 disables it).

Tool calls render as compact cards instead of raw argument JSON. A colored dot leads each card — yellow while running, green on success, red on failure (a shell command with a non-zero exit or signal also turns red, with a [退出码 N] / [信号 X] suffix). Shell tools show the command as $ command, file mutations (edit / write / str_replace_editor) render the applied change git-style: a path header, - lines on a light-red background, + lines on a light-green background, and a └ +N -M · K file(s) footer. File-mutation diffs are shown in full (never collapsed to a … more line) and their cards start expanded. Other tools show a short argument summary and start collapsed to a single line (the command, truncated with … when long); expand to reveal output or the result body. Expanded generic calls convert their JSON arguments and JSON results into readable indented content — key/value fields, bullet lists, and multiline blocks for code/content — instead of raw JSON text.

A web-aligned session stats line sits below the input box: turn/step counts, model and tool wall time, first-token latency, tokens/second, cache-hit percentage, and billed input/output tokens (输入 12.3K · 输出 1.2K), updated as the session progresses.

While a turn is waiting on the provider, the status line shows 等待响应 Ns; if nothing arrives for 60s a warning appears and Esc / Ctrl+C cancels the turn. Follow-ups sent while a turn is running are acknowledged immediately (⚡ … 排队 N) and take effect at the next step boundary, so the UI never looks frozen. Long-running work is not misclassified: while tools are executing the status shows 工具执行中 N, and while subagents are running it shows 子代理执行中 N (no 等待响应/stall warning). Subagent start/end, child assistant output, child tool calls/results, approvals, and ask_user_question prompts are all rendered with a [子代理 …] label; /subagents lists active runs.

/mode opens the agent-mode picker backed by dsh's official preset roster: 标准模式 (standard), PTC 模式 (code), 极简模式 (minimal), 创造模式 (cordis), plus any locally authored presets. On a session that has not produced work the switch applies immediately; otherwise it is remembered as the default for the next launch. The active mode is shown in the header/status line.

/resume switches the running TUI to a past session. With no argument it opens a picker of recent sessions (excluding subagents), labeled by the user's first message with a time/cwd description; /resume switches directly. Switching is refused while a turn is running.

dsh --profile tui --model deepseek-v4-flash
dsh --profile tui --no-color
dsh --profile tui --resume 

dsh --profile tui starts a fresh session directly in the main interface. dsh --profile tui --resume (or dsh --profile tui resume) opens the history-session picker before the main interface; dsh --profile tui --resume (or dsh --profile tui resume ) skips the picker and resumes directly. dsh --profile tui --new explicitly starts fresh without the picker. The in-app /resume command remains available for switching while running.