Completed
on 6 Jan 2026, 10:48 am

Output

gvShell v1 - Initial Features

An interactive AI shell following Unix philosophy - responses become files, prompts become commands, providers compose like pipes.

Core Architecture

REPL Engine Interactive loop with readline, history, auto-completion
Response Manager Store, reference ($1, $2), and extract (.code, .tokens) from responses
Agent Layer Internal infrastructure: providers, SQLite storage, WebSocket to hub
Output Modes Plain text (human) or JSON (RPC mode for automation)

AI Provider Plugins (10 Providers)

/claude  - Anthropic Claude (opus, sonnet, haiku)
/gpt     - OpenAI GPT (gpt-4o, gpt-4-turbo, gpt-3.5)
/gemini  - Google Gemini (pro, ultra, flash)
/ollama  - Local models (llama3, codellama, mistral)
/ask     - Default provider (configurable)
+ 5 more planned providers

Response Reference System

$1              Most recent response
$2, $3...       Previous responses
$claude         Last Claude response
$gpt            Last GPT response
$saved.name     Named saved response
$1.code         Extract code blocks
$1.tokens       Token count
$1.cost         Cost in USD

Session Management

/new       Start fresh session (with project context)
/history   View/search/export session history
/save      Save response with alias
/load      Load saved response
/sessions  List all sessions
/resume    Resume previous session

Authentication & Bootstrap

Three setup paths:

  1. greatVibe.ai sign-in - OAuth device flow ("VIBE-A7K9" code), full features + sync
  2. Local models only - Ollama, offline/privacy-focused
  3. Manual API keys - Use your own Claude/OpenAI keys

gvSystem - Privileged Assistant

Special AI session for configuration and support:

  • Configure providers (enable/disable, set defaults)
  • Troubleshoot connection issues
  • Guided setup for advanced features
  • Cannot access user files (security sandboxed)

Unix Integration

# Pipe between providers
gv> /claude explain quicksort | /gpt summarize this

# Extract and save code
gv> $1.code > sort.py

# Works with grep, sed, awk
gv> /claude list functions | grep "def "

Installation Paths

brew install gvshell              # macOS
apt install gvshell               # Ubuntu/Debian
winget install greatvibe.gvshell  # Windows

What's Deferred (Not in v1)

gvBacklog, gvSprint, gvFlow (workflow management)
gvReview, gvChange (code review objects)
IDE plugins (VS Code, JetBrains)
Enterprise features (SSO, policy engine)
Database plugins (gvshell-postgres, etc.)

Summary

gvShell v1 focuses on the interactive AI shell experience with Claude - response references, multi-provider piping, session management, and hub connectivity. Everything else comes later.

Quick Actions

Original Request

No layout configured

Details

Type General
Status Completed
Scope product
Tags docsgvShellfeaturesroadmap
Created 6 Jan 2026, 10:48 am
Updated 6 Jan 2026, 10:48 am
Timing
⏱️ 73.5s
0 tools
TTFE 1% Thinking 99% Tools 0%
Created By claude

Raw Data

{
  "id": "9a988702-8f3d-4ea7-8d49-a8afefa1c29c",
  "type": "general",
  "status": "completed",
  "title": "gvShell v1 Initial Features Overview",
  "description": "Core features planned for gvShell first release",
  "context": {
    "output": "<div style=\"background:#1e1e1e;color:#e0e0e0;padding:20px;border-radius:8px;font-family:sans-serif;\">\n<h2 style=\"color:#fff;margin:0 0 16px 0;\">gvShell v1 - Initial Features</h2>\n<p style=\"color:#9ca3af;margin:0 0 20px 0;\">An interactive AI shell following Unix philosophy - responses become files, prompts become commands, providers compose like pipes.</p>\n\n<h3 style=\"color:#61afef;margin:20px 0 12px 0;\">Core Architecture</h3>\n<div style=\"background:#2d2d2d;padding:16px;border-radius:6px;margin-bottom:16px;\">\n<table style=\"width:100%;border-collapse:collapse;\">\n<tr style=\"border-bottom:1px solid #444;\">\n  <td style=\"padding:8px;color:#98c379;font-weight:bold;\">REPL Engine</td>\n  <td style=\"padding:8px;color:#e0e0e0;\">Interactive loop with readline, history, auto-completion</td>\n</tr>\n<tr style=\"border-bottom:1px solid #444;\">\n  <td style=\"padding:8px;color:#98c379;font-weight:bold;\">Response Manager</td>\n  <td style=\"padding:8px;color:#e0e0e0;\">Store, reference ($1, $2), and extract (.code, .tokens) from responses</td>\n</tr>\n<tr style=\"border-bottom:1px solid #444;\">\n  <td style=\"padding:8px;color:#98c379;font-weight:bold;\">Agent Layer</td>\n  <td style=\"padding:8px;color:#e0e0e0;\">Internal infrastructure: providers, SQLite storage, WebSocket to hub</td>\n</tr>\n<tr>\n  <td style=\"padding:8px;color:#98c379;font-weight:bold;\">Output Modes</td>\n  <td style=\"padding:8px;color:#e0e0e0;\">Plain text (human) or JSON (RPC mode for automation)</td>\n</tr>\n</table>\n</div>\n\n<h3 style=\"color:#61afef;margin:20px 0 12px 0;\">AI Provider Plugins (10 Providers)</h3>\n<div style=\"background:#2d2d2d;padding:16px;border-radius:6px;margin-bottom:16px;\">\n<pre style=\"margin:0;color:#e0e0e0;line-height:1.8;\">\n<span style=\"color:#c678dd;\">/claude</span>  - Anthropic Claude (opus, sonnet, haiku)\n<span style=\"color:#c678dd;\">/gpt</span>     - OpenAI GPT (gpt-4o, gpt-4-turbo, gpt-3.5)\n<span style=\"color:#c678dd;\">/gemini</span>  - Google Gemini (pro, ultra, flash)\n<span style=\"color:#c678dd;\">/ollama</span>  - Local models (llama3, codellama, mistral)\n<span style=\"color:#c678dd;\">/ask</span>     - Default provider (configurable)\n<span style=\"color:#7f848e;\">+ 5 more planned providers</span>\n</pre>\n</div>\n\n<h3 style=\"color:#61afef;margin:20px 0 12px 0;\">Response Reference System</h3>\n<div style=\"background:#2d2d2d;padding:16px;border-radius:6px;margin-bottom:16px;\">\n<pre style=\"margin:0;color:#e0e0e0;font-family:monospace;line-height:1.8;\">\n<span style=\"color:#e5c07b;\">$1</span>              Most recent response\n<span style=\"color:#e5c07b;\">$2</span>, <span style=\"color:#e5c07b;\">$3</span>...       Previous responses\n<span style=\"color:#e5c07b;\">$claude</span>         Last Claude response\n<span style=\"color:#e5c07b;\">$gpt</span>            Last GPT response\n<span style=\"color:#e5c07b;\">$saved.name</span>     Named saved response\n<span style=\"color:#e5c07b;\">$1.code</span>         Extract code blocks\n<span style=\"color:#e5c07b;\">$1.tokens</span>       Token count\n<span style=\"color:#e5c07b;\">$1.cost</span>         Cost in USD\n</pre>\n</div>\n\n<h3 style=\"color:#61afef;margin:20px 0 12px 0;\">Session Management</h3>\n<div style=\"background:#2d2d2d;padding:16px;border-radius:6px;margin-bottom:16px;\">\n<pre style=\"margin:0;color:#e0e0e0;font-family:monospace;line-height:1.8;\">\n<span style=\"color:#c678dd;\">/new</span>       Start fresh session (with project context)\n<span style=\"color:#c678dd;\">/history</span>   View/search/export session history\n<span style=\"color:#c678dd;\">/save</span>      Save response with alias\n<span style=\"color:#c678dd;\">/load</span>      Load saved response\n<span style=\"color:#c678dd;\">/sessions</span>  List all sessions\n<span style=\"color:#c678dd;\">/resume</span>    Resume previous session\n</pre>\n</div>\n\n<h3 style=\"color:#61afef;margin:20px 0 12px 0;\">Authentication & Bootstrap</h3>\n<div style=\"background:#2d2d2d;padding:16px;border-radius:6px;margin-bottom:16px;\">\n<p style=\"margin:0 0 12px 0;\"><strong style=\"color:#98c379;\">Three setup paths:</strong></p>\n<ol style=\"margin:0;padding-left:20px;color:#e0e0e0;line-height:1.8;\">\n<li><strong>greatVibe.ai sign-in</strong> - OAuth device flow (\"VIBE-A7K9\" code), full features + sync</li>\n<li><strong>Local models only</strong> - Ollama, offline/privacy-focused</li>\n<li><strong>Manual API keys</strong> - Use your own Claude/OpenAI keys</li>\n</ol>\n</div>\n\n<h3 style=\"color:#61afef;margin:20px 0 12px 0;\">gvSystem - Privileged Assistant</h3>\n<div style=\"background:#2d2d2d;padding:16px;border-radius:6px;margin-bottom:16px;\">\n<p style=\"margin:0;color:#e0e0e0;\">Special AI session for configuration and support:</p>\n<ul style=\"margin:8px 0 0 0;padding-left:20px;color:#9ca3af;line-height:1.8;\">\n<li>Configure providers (enable/disable, set defaults)</li>\n<li>Troubleshoot connection issues</li>\n<li>Guided setup for advanced features</li>\n<li>Cannot access user files (security sandboxed)</li>\n</ul>\n</div>\n\n<h3 style=\"color:#61afef;margin:20px 0 12px 0;\">Unix Integration</h3>\n<div style=\"background:#2d2d2d;padding:16px;border-radius:6px;margin-bottom:16px;\">\n<pre style=\"margin:0;color:#e0e0e0;font-family:monospace;line-height:1.8;\">\n<span style=\"color:#7f848e;\"># Pipe between providers</span>\ngv> /claude explain quicksort | /gpt summarize this\n\n<span style=\"color:#7f848e;\"># Extract and save code</span>\ngv> $1.code > sort.py\n\n<span style=\"color:#7f848e;\"># Works with grep, sed, awk</span>\ngv> /claude list functions | grep \"def \"\n</pre>\n</div>\n\n<h3 style=\"color:#61afef;margin:20px 0 12px 0;\">Installation Paths</h3>\n<div style=\"background:#2d2d2d;padding:16px;border-radius:6px;margin-bottom:16px;\">\n<pre style=\"margin:0;color:#e0e0e0;font-family:monospace;line-height:1.6;\">\nbrew install gvshell              <span style=\"color:#7f848e;\"># macOS</span>\napt install gvshell               <span style=\"color:#7f848e;\"># Ubuntu/Debian</span>\nwinget install greatvibe.gvshell  <span style=\"color:#7f848e;\"># Windows</span>\n</pre>\n</div>\n\n<h3 style=\"color:#61afef;margin:20px 0 12px 0;\">What's Deferred (Not in v1)</h3>\n<div style=\"background:#2d2d2d;padding:16px;border-radius:6px;margin-bottom:16px;\">\n<p style=\"margin:0;color:#9ca3af;line-height:1.8;\">\n<span style=\"color:#e06c75;\">✗</span> gvBacklog, gvSprint, gvFlow (workflow management)<br>\n<span style=\"color:#e06c75;\">✗</span> gvReview, gvChange (code review objects)<br>\n<span style=\"color:#e06c75;\">✗</span> IDE plugins (VS Code, JetBrains)<br>\n<span style=\"color:#e06c75;\">✗</span> Enterprise features (SSO, policy engine)<br>\n<span style=\"color:#e06c75;\">✗</span> Database plugins (gvshell-postgres, etc.)\n</p>\n</div>\n\n<div style=\"background:#2d2d2d;padding:16px;border-radius:6px;margin-top:20px;\">\n<h4 style=\"margin:0 0 8px 0;color:#fff;\">Summary</h4>\n<p style=\"color:#9ca3af;margin:0;\">gvShell v1 focuses on the <strong style=\"color:#98c379;\">interactive AI shell experience with Claude</strong> - response references, multi-provider piping, session management, and hub connectivity. Everything else comes later.</p>\n</div>\n</div>",
    "choices": [
      {
        "label": "Read full design",
        "value": "Show me the complete gvShellDesign.md document",
        "primary": true
      },
      {
        "label": "View project structure",
        "value": "Show me the gvShell project structure for Go development"
      },
      {
        "label": "Start implementation",
        "value": "What's the first step to start building gvShell?"
      }
    ],
    "requestId": "f8d5c14d-7f11-4b69-b598-cc2737259537",
    "turnTiming": {
      "totalMs": 73456,
      "ttfeMs": 514,
      "thinkingMs": 72942,
      "toolExecutionMs": 0,
      "toolCallCount": 0,
      "thinkingPct": 99,
      "toolsPct": 0,
      "ttfePct": 1
    }
  },
  "createdBy": "claude",
  "createdAt": "2026-01-06T00:48:31.828Z",
  "updatedAt": "2026-01-06T00:48:38.920Z",
  "requestId": "f8d5c14d-7f11-4b69-b598-cc2737259537",
  "scope": "product",
  "tags": [
    "docs",
    "gvShell",
    "features",
    "roadmap"
  ],
  "targetUser": "claude"
}
DashboardReportsKontasksSessionsTelemetryLogs + Go