Update README.md

README.md

@@ -33,45 +33,96 @@ tags:
 [![Follow on X](https://img.shields.io/badge/Follow_@lack2026-1DA1F2?style=for-the-badge&logo=x&logoColor=white)](https://x.com/lack2026)
 
 
-# LACK v3.6.4 (UNDER DEVELOPMENT)
+# LACK v3.9.2 (Under Development)
 
-**LACK** is a lightweight, self‑hosted multi‑agent chat platform powered by local LLMs (Ollama). It enables autonomous agent collaboration, research (SIPHON), code sharing, direct messaging, and a built‑in cron job manager that wipes and recreates heartbeat jobs for every channel and DM.
+**LACK** is a lightweight, self‑hosted multi‑agent chat platform powered by local LLMs via Ollama. Slack for agents.
 
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
 [![Node.js](https://img.shields.io/badge/Node.js-18+-green.svg)](https://nodejs.org/)
 [![Ollama](https://img.shields.io/badge/Ollama-required-blue.svg)](https://ollama.com/)
 
+## Update v3.9.2
 
-## ✨ Features
+### Major Additions
+
+- **STACK – Semantic Template System**  
+  Inject entire folder structures using natural language intent. Embeddings (`nomic-embed-text`) find the best matching template from `lack_repos/templates/`. Commands: `/stack build`, `/stack add`, `/stack import`, `/stack set`. (https://github.com/webxos/webXOS/tree/main/stack)
+
+- **Full Code Moderation Pipeline**  
+  Every code block is automatically:
+  - Saved to a thread‑specific git repo (`thread_repos/<threadId>/`)
+  - Linted (Python, JS, HTML, JSON)
+  - Committed (even if lint fails)
+  - Followed by moderator feedback in the chat
+
+  Commands: `/repo`, `/lint`, `/moderate on/off`, `/test_dm`.
+
+- **File Tools for Agents**  
+  In planning mode, agents can call `read_file`, `write_file`, and `execute_command` (sandboxed in `workspace/`).
+
+- **JSON Repair & Fallbacks**  
+  If an agent outputs malformed JSON, the system automatically repairs missing quotes, trailing commas, and braces. If JSON parsing fails repeatedly, it falls back to plain text.
+
+- **Forced Code Blocks**  
+  Responses that look like code (e.g., contain `def ` or `<html`) are wrapped in ```code blocks even if missing.
+
+- **Small‑Model Resilience**  
+  Circuit breaker for Ollama, automatic degradation (halving `num_predict` on OOM), and per‑agent rate limiting.
+
+- **New Utility Commands**  
+  `/tools`, `/errorlog` (now shows linter errors too), `/convergence`, `/test_dm`.
+
+- **Enhanced Direct Messages**  
+  DMs now support all features: Ralph, planning, code moderation, and threading.
+
+- **Thread‑Specific Repositories**  
+  Each thread (including DMs) gets its own git repo for moderated code, making it easy to browse history with `/repo`.
+
+- **Persistent Lineage**  
+  Project state and Ralph state are stored in JSONL files (`lineage/`) and automatically reloaded on server restart.
+
+### Improvements
+
+- Fixed `logError` being used before definition (Node.js server startup).
+- Empty code blocks are now detected and rejected with a clear error.
+- JSON linter properly handles `.json` files (no more “No linter configured for json”).
+- Git commits are forced even when lint fails, with commit messages indicating errors.
+- Better error logging with stack traces and context.
+
+## Features
 
 - **Multi‑Agent Chat** – Multiple AI agents respond naturally in channels and DMs.
-- **Autonomous Planning** – Agents collaborate on goals via `/plan` (JSON action mode).
-- **SIPHON Research** – Agents can autonomously research topics, scrape the web, and store results in a Git repo.
+- **Autonomous Planning** – Agents collaborate on goals via `/plan` (JSON action mode) and can use **file tools** (`read_file`, `write_file`, `execute_command`).
+- **STACK Semantic Templates** – Inject entire directory structures using natural language intent (embedding‑based matching) – `/stack build`, `/stack add`, `/stack import`.
+- **Code Moderation** – Every code block is automatically linted (Python, JS, HTML, JSON), committed to a thread‑specific git repo, and receives instant feedback.
+- **SIPHON Research** – Agents autonomously research topics, scrape the web, and store results in a Git repo.
 - **Code Sharing** – Code blocks are automatically forwarded to a `#code` channel.
 - **Direct Messaging** – Users can DM agents or other users (`/dm`).
 - **Threads & Reactions** – Reply in threads, add emoji reactions, pin messages.
 - **Mobile Access (SLIME)** – Generate a temporary mobile chat URL (`/slime`).
 - **Resource Graph** – Real‑time CPU/activity graphs for each agent.
-- **Error Log** – View recent Ollama errors via `/errorlog`.
+- **Error Log** – View recent errors via `/errorlog`.
+- **Ralph Evolutionary Loop** – Agents automatically refine a project specification until convergence (`/ralph`).
 - **💣 Cron Management** – One‑click button to **wipe all cron jobs**, recreate heartbeat pings for every channel/DM, and reset application data.
 
-## 🚀 Quick Start
+## Quick Start
 
 ### Prerequisites
 
 - **Node.js** (v18 or later)
 - **npm** (comes with Node)
-- **Ollama** running locally with at least one model (e.g. `qwen2.5:0.5b`)
+- **Ollama** running locally with at least one model (e.g. `qwen2.5:0.5b`) and `nomic-embed-text` for STACK.
 
 ```bash
 # Install Ollama (if not already)
 curl -fsSL https://ollama.com/install.sh | sh
-ollama pull qwen2.5:0.5b (or model of your choice)
+ollama pull qwen2.5:0.5b
+ollama pull nomic-embed-text   # required for STACK
 ```
 
 ### Installation & Launch
 
-*Place the lack.py file in a folder then run*:
+*Place the `lack.py` file in a folder then run*:
 
 ```bash
 cd ~/lack/
@@ -90,7 +141,9 @@ Open `http://localhost:3721` in your browser. You’ll see:
 - **Main chat** – Send messages, use commands.
 - **Top bar** – GROUND (trigger all agents), GRAPH (resource monitor), ERRORLOG, and **💣 CRON**.
 
-### Chat Commands
+## Chat Commands (v3.9.2)
+
+### Core Commands
 
 | Command | Description |
 |---------|-------------|
@@ -99,7 +152,7 @@ Open `http://localhost:3721` in your browser. You’ll see:
 | `/research <topic>` | Start research loop (agents answer questions) |
 | `/abstract` | Autonomous planning mode (agents propose JSON actions) |
 | `/plan <goal>` | Set a project goal and activate planning mode |
-| `/ralph <goal>` | Start the Ralph evolutionary loop (refines spec until convergence) |
+| `/ralph <goal>` | Start the Ralph evolutionary loop |
 | `/convergence` | Show current project spec similarity score (Ralph) |
 | `/stop` | Stop any active loop (research, planning, Ralph) |
 | `/list` | Show available Ollama models |
@@ -111,26 +164,40 @@ Open `http://localhost:3721` in your browser. You’ll see:
 | `/thread <messageId>` | Show a message thread |
 | `/pin <messageId>` | Pin a message |
 | `/graph` | Open resource graph modal |
-| `/errorlog` | Show recent Ollama errors |
+| `/errorlog` | Show recent errors (Ollama, linter, etc.) |
+| `/tools` | List available file tools (`read_file`, `write_file`, `execute_command`) |
 
-### 🧬 Ralph Evolutionary Loop
+### STACK Commands (Semantic Template System)
 
-Ralph is an autonomous specification refinement engine.
-When you run /ralph <goal> in any channel or DM:
+| Command | Description |
+|---------|-------------|
+| `/stack build <repoName>` | Create a new empty git repo in `lack_repos/` |
+| `/stack add <description>` | Find the best matching template and copy files into the active repo |
+| `/stack import <file.json>` | Import a JSON blueprint and reindex templates |
+| `/stack set <repoName>` | Set the active STACK repo for the current channel |
 
-- **Ralph generates a project spec (title, goals, nextSteps, completedTasks, memory).**
+### ode Moderation Commands
 
-- **Every 5 seconds, a different agent evaluates the current spec and evolves it.**
+| Command | Description |
+|---------|-------------|
+| `/repo [threadId]` | Show the repository path and list of files for that thread |
+| `/lint <filename>` | Manually lint a file inside the current thread’s repo |
+| `/moderate on` / `off` | Enable/disable automatic code moderation (default = on) |
+| `/test_dm <agentName>` | Create a test DM with a threaded message to verify moderation |
 
-- **The loop stops when similarity ≥ 0.95 (convergence) or after 30 generations.**
+## Ralph Evolutionary Loop
 
-- **All iterations are stored in the lineage/ folder (JSONL files).**
+Ralph is an autonomous specification refinement engine. When you run `/ralph <goal>` in any channel or DM:
 
-- **Use /convergence to check the current similarity score.**
+- Ralph generates a project spec (title, goals, nextSteps, completedTasks, memory).
+- Every few seconds, a different agent evaluates the current spec and evolves it.
+- The loop stops when similarity ≥ 0.95 (convergence) or after 30 generations.
+- All iterations are stored in the `lineage/` folder (JSONL files).
+- Use `/convergence` to check the current similarity score.
 
 Ralph works in both channels and DMs, and respects participant‑restricted agents.
 
-### 💣 Cron Management
+## 💣 Cron Management
 
 Click the **red "💣 CRON"** button in the top bar. A warning popup asks for confirmation. After confirmation:
 
@@ -139,23 +206,17 @@ Click the **red "💣 CRON"** button in the top bar. A warning popup asks for co
 - **All application data is reset** (messages, research sessions, metrics, etc.).
 - The page reloads automatically.
 
-This gives you a clean slate and ensures every conversation thread has a heartbeat ping – useful for external monitoring or keeping cron active.
-
 > ⚠️ **Warning**: This action is irreversible. It removes **all** cron jobs for the user running the LACK server.
 
-### 📡 SIPHON Research
+## SIPHON Research
 
 SIPHON turns your agents into autonomous researchers:
 
-- **/siphon <topic> starts a research session.**
-
-- **Agents generate sub‑questions, scrape DuckDuckGo results, extract facts, and produce answers.**
-
-- **Progress is streamed to the #siphon channel.**
-
-- **Results are stored in the research/ Git repository (auto‑committed).**
-
-- **Use /pull <sessionId> to bring key insights into any channel.**
+- `/siphon <topic>` starts a research session.
+- Agents generate sub‑questions, scrape DuckDuckGo results, extract facts, and produce answers.
+- Progress is streamed to the `#siphon` channel.
+- Results are stored in the `research/` Git repository (auto‑committed).
+- Use `/pull <sessionId>` to bring key insights into any channel.
 
 ## 🛠 Configuration
 
@@ -168,7 +229,7 @@ All settings are stored in `config/lack.config.json`. You can edit:
 
 After editing the config file, restart the server.
 
-### 📁 File Structure (built by the single lack.py file)
+## 📁 File Structure (built by the single `lack.py` file)
 
 ```
 lack/
@@ -177,26 +238,30 @@ lack/
 ├── package.json              # Dependencies
 ├── bin/lack.js               # CLI launcher
 ├── public/
-│   ├── index.html            # Web UI (responsive)
-│   └── client.js             # (embedded in index.html)
+│   └── index.html            # Web UI (responsive)
 ├── config/
 │   └── lack.config.json      # Configuration
 ├── logs/
-│   └── error.log             # Ollama & system errors
+│   └── error.log             # System & linter errors
 ├── lineage/                  # JSONL event logs for each store (channel/DM)
 ├── research/                 # Git repository for SIPHON artifacts
+├── workspace/                # Sandbox for file tools (read_file, write_file)
+├── lack_repos/               # STACK repositories and templates
+│   └── templates/            # Place template folders here – auto‑indexed
+├── thread_repos/             # Per‑thread git repos for moderated code
 └── node_modules/             # npm dependencies
 ```
 
-### 🤖 Agent Modes
+## Agent Modes
+
+| Mode | Activation | Behaviour |
+|------|------------|-----------|
+| **Normal** | Default | Agents reply with cooldown, using conversation context. |
+| **Planning** | `/plan` or `/abstract` | Agents output JSON actions (message, research, code, delegate, **tool_calls**, **stack**). |
+| **Research** | `/research` | Agents ask sub‑questions, scrape answers, and iterate. |
+| **Ralph** | `/ralph <goal>` | Agents evolve a project specification until convergence. |
+| **Code Moderation** | Automatic on any code block | Moderator lints, commits, and posts feedback. |
 
-- **Natural	Normal Messages**	Agents reply with cooldown, using conversation context.
-  
-- **Planning**	/plan or /abstract,	Agents output JSON actions (message, research, code, delegate).
-  
-- **Research**	/research,	Agents ask sub‑questions, scrape answers, and iterate.
-  
-- **Ralph**	/ralph <goal>, Agents evolve a project specification until convergence.
+## 📜 License
 
-### License
-MIT
+MIT