aiden-runtime 3.16.2 → 3.18.0

package/README.md

@@ -1,14 +1,18 @@
 ```
- █████╗ ██╗██████╗ ███████╗███╗   ██╗
+█████╗  ██╗██████╗ ███████╗███╗   ██╗
 ██╔══██╗██║██╔══██╗██╔════╝████╗  ██║
 ███████║██║██║  ██║█████╗  ██╔██╗ ██║
 ██╔══██║██║██║  ██║██╔══╝  ██║╚██╗██║
 ██║  ██║██║██████╔╝███████╗██║ ╚████║
 ╚═╝  ╚═╝╚═╝╚═════╝ ╚══════╝╚═╝  ╚═══╝
 
-local-first AI operating system
-1,400+ skills · 80+ tools · 15+ providers · AGPL-3.0
-Windows · Linux · WSL · macOS (API mode)
+Autonomous AI Operating System
+
+1,500+ Skills • 89+ Tools • 14+ Providers • AGPL-3.0
+
+Windows • Linux • WSL • macOS (API Mode)
+
+Self-Healing • Browser Automation • Terminal Control • Persistent Memory
 ```
 
 <p align="center">
@@ -17,6 +21,7 @@ Windows · Linux · WSL · macOS (API mode)
   <a href="https://discord.gg/gMZ3hUnQTm"><img src="https://img.shields.io/badge/chat-discord-7289da?logo=discord&logoColor=white" alt="Discord" /></a>
   <a href="./LICENSE"><img src="https://img.shields.io/badge/license-AGPL--3.0-orange" alt="License: AGPL-3.0" /></a>
   <a href="https://github.com/taracodlabs/aiden/stargazers"><img src="https://img.shields.io/github/stars/taracodlabs/aiden?style=flat&color=f9d71c" alt="Stars" /></a>
+  <a href="https://www.npmjs.com/package/aiden-os"><img src="https://img.shields.io/npm/v/aiden-os?color=f97316&label=npm" alt="npm" /></a>
 </p>
 
 <p align="center">
@@ -28,8 +33,21 @@ Windows · Linux · WSL · macOS (API mode)
 
 ---
 
-> **v3.16 — `npx aiden-os` single-command launch · programmatic API · in-process server + CLI**
-> Run Aiden with one command — no clone, no build, no two terminals. `npx aiden-os` installs the runtime, runs the setup wizard on first use, then starts the server and CLI together in the same process. New `start()` / `run()` exports for embedding Aiden in your own Node.js apps. See [changelog](#changelog) below.
+> **v3.18 — live dropdown UX · real PC control · smart model failover · anti-confabulation**
+> Type `/` for 63 commands or `@` for 61 tools with instant dropdown. Open/close apps, change volume, and control your PC for real — no more fake responses. Smart per-model failover with free-tier defaults. See [changelog](#changelog) below.
+
+---
+
+## Why Aiden
+
+Most AI agents answer questions. Aiden executes work.
+
+- **Runs on your machine** — local-first, no telemetry, no cloud required
+- **Controls your desktop** — vision loop, mouse, keyboard, window management
+- **Automates any browser** — navigate, click, extract via playwright-cli
+- **Learns from every session** — writes skills from successes, lessons from failures
+- **Works fully offline** — Ollama support, zero cloud dependency
+- **One command to start** — `npx aiden-os` installs, configures, runs everything
 
 ---
 
@@ -192,7 +210,7 @@ Model:     aiden-3.13
 
 ![TUI](docs/images/tui.png)
 
-Full command palette, 1,400+ skills, 80+ tools, automatic provider routing (Groq → BOA → Ollama). Runs in any terminal.
+Full command palette, 1,400+ skills, 89+ tools, automatic provider routing (Groq → OpenRouter → Ollama). Runs in any terminal.
 
 ### Desktop app
 
@@ -223,6 +241,23 @@ Full chat interface with live activity panel. Local-first, connects to Ollama or
 
 ---
 
+## Aiden vs alternatives
+
+| Feature | Aiden | Hermes | OpenClaw |
+|---|---|---|---|
+| **Windows native installer** | ✅ | ❌ | ❌ |
+| **Desktop OS control** | ✅ vision + mouse + keyboard | ❌ | ❌ |
+| **One-command install** | ✅ `npx aiden-os` | ❌ | ❌ |
+| **OpenAI-compatible API** | ✅ `/v1/chat/completions` | ❌ | ❌ |
+| **agentskills.io skills** | ✅ 1500+ | ✅ | ✅ 13K+ |
+| **Offline (Ollama)** | ✅ | ✅ | ✅ |
+| **Local web dashboard** | ✅ | ✅ | ✅ |
+| **Pro license system** | ✅ | ❌ | ❌ |
+| **Zero CVEs** | ✅ | ❌ | ❌ |
+| **License** | AGPL-3.0 | MIT | MIT |
+
+---
+
 ## Architecture
 
 ```
@@ -353,6 +388,42 @@ Or toggle live from the Aiden CLI without restarting:
 
 ---
 
+## Commands
+
+### Start Aiden
+| Command | Description |
+|---------|-------------|
+| `npx aiden-os` | Install, configure, and start (recommended) |
+| `npm start` | Start API server (port 4200) |
+| `npm run cli` | Start TUI (connect to running server) |
+| `npm run build` | Rebuild after source changes |
+| `aiden --reconfigure` | Re-run setup wizard, change providers |
+
+### In-chat commands
+| Command | Description |
+|---------|-------------|
+| `/help` | Show all commands |
+| `/switch <provider>` | Change primary provider live |
+| `/budget` | Show daily token spend + remaining |
+| `/budget set <n>` | Set daily limit in USD |
+| `/memory` | View distilled facts and memory stats |
+| `/memory search <q>` | Search remembered facts |
+| `/profile` | View structured user profile |
+| `/failed [reason]` | Teach Aiden from a wrong answer |
+| `/skills` | List loaded skills |
+| `/install <skill>` | Install from community registry |
+| `/publish <skill>` | Publish skill to registry |
+| `/skills validate <n>` | Validate agentskills.io compliance |
+| `/sandbox status` | Docker sandbox mode |
+| `/sandbox auto` | Enable sandboxed shell/python |
+| `/permissions` | View permission mode |
+| `/permissions ask` | Require approval for destructive ops |
+| `/permissions allow` | Allow all operations silently |
+| `/retry` | Retry last query |
+| `/exit` | Save memory and exit |
+
+---
+
 ## Contributing
 
 Contributions are welcome — see [CONTRIBUTING.md](CONTRIBUTING.md) for the full guide.
@@ -376,6 +447,103 @@ Contributions are welcome — see [CONTRIBUTING.md](CONTRIBUTING.md) for the ful
 
 ## Changelog
 
+### v3.18.0 — 2026-04-30
+
+**Live dropdown UX (Hermes-style)**
+- Type `/` for instant command dropdown (63 commands)
+- Type `@` for tool dropdown (61 tools)
+- Prefix-match filter, arrow nav, Tab to select, Esc to close
+
+**Real PC control**
+- `close chrome` / `spotify` / `notepad` → actually closes via taskkill
+- `increase/decrease volume by N` → actually changes
+- `mute/unmute` → actually toggles
+- 30+ app name → exe map
+- `system_volume` detects intent from any natural input
+
+**YouTube auto-plays**
+- `play X on youtube` → opens browser, auto-clicks first result
+- Bypasses fast-path that was blocking it
+
+**Anti-confabulation rules**
+- SOUL.md updated: never claim actions completed without tool calls
+- InstantAction shortcuts that faked actions removed
+- Honest fallback messages when providers fail
+
+**Smart provider failover**
+- 3-strike rule: provider disabled for 15 min after 3 rate limits
+- Permanent disable on 401/403 (invalid key)
+- All cloud failed → automatic Ollama fallback
+
+**Smart model selection**
+- Free tier defaults per provider (Llama 70B free, Gemini 2.5 Flash, etc.)
+- Per-model failover within provider before marking provider rate-limited
+- Override with `PROVIDER_MODEL` env var
+- `/models` command shows per-provider table with FREE/PAID badges
+
+**Server logs no longer leak into chat**
+- `console.log` redirected to stderr
+- CLI output is clean even with both in same terminal
+
+**Skill loader fix**
+- 1,484 skills now load (was blocking 1,445 due to overly broad patterns)
+
+**Known issues — fixing in v3.19**
+- Cross-provider failover not always reliable (Groq may not try other providers)
+- Real-time state queries (now playing, open tabs) need dedicated tools
+- YouTube auto-click occasionally fails on slow-loading pages
+
+---
+
+### v3.17.0 — 2026-04-28
+
+**Local web dashboard**
+- Browser UI at localhost:4200/ui — no terminal needed
+- Chat, Providers, Memory, Skills panels
+- Live SSE streaming, markdown rendering
+
+**Plugin system**
+- Drop workspace/plugins/<name>.js → auto-loads
+- preTool/postTool hooks, custom tool registration
+- Session lifecycle hooks (onSessionStart/onSessionEnd)
+- Hot-reload with /plugins reload
+- Examples: audit-log.js, hello-tool.js
+
+**MCP server mode**
+- Expose Aiden's tools to Claude Code, Cursor, VS Code
+- Run: node dist-bundle/cli.js mcp
+- 28 safe tools exposed by default
+- MCP_ALLOW_DESTRUCTIVE=true for full tool access
+- Config examples in docs/mcp/
+
+**Bug fixes**
+- Dashboard chat showed "(no response)" for every message — SSE event field names in the browser client (`ev.type`) didn't match the server's wire format (`ev.token`, `ev.done`, `ev.tool`). All event handlers rewritten to match actual shapes.
+- SOUL.md provider honesty: removed stale BayOfAssets reference, added explicit rule against claiming Ollama when running on Groq/OpenRouter.
+
+---
+
+### v3.16.0 — 2026-04-28
+
+**One-command install**
+- `npx aiden-os` — zero-install launcher; works on Windows, macOS, Linux (Node.js 18+)
+- `aiden-os` npm package bootstraps `aiden-runtime` automatically, no git clone needed
+- Setup wizard on first run with `--reconfigure` flag to re-run anytime
+
+**Security**
+- Shell blocklist — dangerous commands flagged before execution
+- Permission mode — explicit user approval gate for destructive actions
+- Token budget enforcement — per-request ceiling to prevent runaway loops
+
+**Memory**
+- Conflict resolution — contradictory memories detected and reconciled automatically
+- `/memory` command — inspect, edit, and prune the memory store from the CLI
+
+**UX**
+- Aiden branded banner replaces DevOS; orange `#FF6B35` identity throughout CLI
+- `--reconfigure` flag to re-run first-time setup without reinstalling
+
+---
+
 ### v3.15.0 — 2026-04-28
 
 **Browser & Automation**
@@ -460,6 +628,16 @@ See [releases page](https://github.com/taracodlabs/aiden-releases/releases) for
 
 ---
 
+## Sponsors
+
+Aiden is built and maintained by one person.
+If it saves you time, consider sponsoring:
+
+[![Sponsor on GitHub](https://img.shields.io/badge/sponsor-❤-pink)](https://github.com/sponsors/shivadeore111-design)
+[![Donate via Razorpay](https://img.shields.io/badge/donate-Razorpay-blue)](https://razorpay.me/@shivadeore)
+
+---
+
 ## License
 
 | Component | License |

package/config/devos.config.json

@@ -14,9 +14,9 @@
     "apis": [
       {
         "model": "gemini-3-flash",
-        "enabled": true,
+        "enabled": false,
         "rateLimited": false,
-        "usageCount": 59,
+        "usageCount": 0,
         "name": "boa-1",
         "provider": "boa",
         "key": "boa-5a949b90-094c-4f35-a6ae-3fab8acc42aa"
@@ -25,7 +25,7 @@
         "model": "llama-3.3-70b-versatile",
         "enabled": true,
         "rateLimited": false,
-        "usageCount": 30,
+        "usageCount": 63,
         "name": "groq-1",
         "provider": "groq",
         "key": "env:GROQ_API_KEY"
@@ -34,7 +34,7 @@
         "model": "llama-3.3-70b-versatile",
         "enabled": true,
         "rateLimited": false,
-        "usageCount": 61,
+        "usageCount": 79,
         "name": "groq-2",
         "provider": "groq",
         "key": "env:GROQ_API_KEY_2"
@@ -43,7 +43,7 @@
         "model": "llama-3.3-70b-versatile",
         "enabled": true,
         "rateLimited": false,
-        "usageCount": 35,
+        "usageCount": 37,
         "name": "groq-3",
         "provider": "groq",
         "key": "env:GROQ_API_KEY_3"
@@ -52,7 +52,7 @@
         "model": "llama-3.3-70b-versatile",
         "enabled": true,
         "rateLimited": false,
-        "usageCount": 10,
+        "usageCount": 11,
         "name": "groq-4",
         "provider": "groq",
         "key": "env:GROQ_API_KEY_4"
@@ -94,46 +94,56 @@
         "key": "env:GEMINI_API_KEY_4"
       },
       {
-        "model": "anthropic/claude-3.5-haiku",
+        "model": "meta-llama/llama-3.3-70b-instruct:free",
         "enabled": true,
         "rateLimited": false,
-        "usageCount": 15,
+        "usageCount": 0,
         "name": "openrouter-1",
         "provider": "openrouter",
         "key": "env:OPENROUTER_API_KEY"
       },
       {
-        "model": "anthropic/claude-3.5-haiku",
+        "model": "meta-llama/llama-3.3-70b-instruct:free",
         "enabled": true,
         "rateLimited": false,
-        "usageCount": 9,
+        "usageCount": 0,
         "name": "openrouter-2",
         "provider": "openrouter",
         "key": "env:OPENROUTER_API_KEY_2"
       },
       {
-        "model": "anthropic/claude-3.5-haiku",
+        "model": "meta-llama/llama-3.3-70b-instruct:free",
         "enabled": true,
         "rateLimited": false,
-        "usageCount": 9,
+        "usageCount": 0,
         "name": "openrouter-3",
         "provider": "openrouter",
         "key": "env:OPENROUTER_API_KEY_3"
       },
       {
-        "model": "anthropic/claude-3.5-haiku",
+        "model": "meta-llama/llama-3.1-405b-instruct",
         "enabled": true,
         "rateLimited": false,
-        "usageCount": 8,
-        "name": "openrouter-4",
-        "provider": "openrouter",
-        "key": "env:OPENROUTER_API_KEY_4"
+        "usageCount": 0,
+        "name": "together-1",
+        "provider": "custom",
+        "baseUrl": "https://api.together.xyz/v1/chat/completions",
+        "key": "env:TOGETHER_API_KEY"
+      },
+      {
+        "model": "nvidia/llama-3.3-nemotron-super-49b-v1",
+        "enabled": true,
+        "rateLimited": false,
+        "usageCount": 0,
+        "name": "nvidia-1",
+        "provider": "nvidia",
+        "key": "env:NVIDIA_API_KEY"
       },
       {
         "model": "gpt-4o-mini",
         "enabled": false,
         "rateLimited": false,
-        "usageCount": 1,
+        "usageCount": 0,
         "name": "boa-2",
         "provider": "boa",
         "key": "boa-5a949b90-094c-4f35-a6ae-3fab8acc42aa"
@@ -182,5 +192,5 @@
   "cli": {
     "theme": "ember"
   },
-  "primaryProvider": "groq-1"
+  "primaryProvider": "groq-2"
 }

package/config/hardware.json

@@ -1,9 +1,9 @@
 {
   "gpu": "NVIDIA GeForce GTX 1060 6GB",
   "vramGB": 6,
-  "ramGB": 16,
+  "ramGB": 32,
   "platform": "windows",
   "cudaAvailable": true,
   "appleSilicon": false,
-  "detectedAt": "2026-04-27T17:24:32.210Z"
+  "detectedAt": "2026-04-29T18:49:39.763Z"
 }