Viden — Development skill for Claude Code

Viden

Viden is a local-first coding-agent cockpit for developers who want one terminal surface to chat with a model, approve real workspace changes, supervise delegated agents, and keep enough evidence to resume work later.

Chinese version: [README.zh-CN.md](README.zh-CN.md)


Why It Exists

Most coding agents are good at a single conversation. Viden is built around a slightly different job: coordinating programming work. It keeps the active conversation, tool effects, approvals, tests, tasks, memory, provider health, and external agent lanes visible in one operator cockpit.

Highlights

• Cockpit TUI: transcript, approval state, workspace snapshot, active tasks, diagnostics, provider health, and recent evidence stay visible together.
• Real tool execution: file read/write/edit, search, shell, web, Git, LSP, test commands, task state, and memory all run through the shared runtime path.
• Permission-aware edits: mutating file, shell, Git, workflow, and delegated agent actions are mediated by permission modes before they affect the workspace.
• Multi-provider runtime: DeepSeek, OpenAI, Anthropic, OpenAI-compatible gateways, Ollama, and offline fallback are available through one provider interface, with provider registry diagnostics.
• Agent supervision lanes: Codex, Claude, custom shell/template lanes, tmux, embedded PTY, and experimental ACP surfaces are exposed as supervised operator lanes.
• Durable local context: transcripts, session index, task events, memory events, lane artifacts, and preview evidence are stored locally by default.
• Screenshot-gated TUI iteration: visual changes produce deterministic screenshots for review before release.
• Multi-platform install: Homebrew plus release archives for macOS, Linux, and Windows.

Screenshots

These are real macOS Terminal screenshots captured from Viden `0.1.30` release-preview states with `deepseek / deep