OpenClaw - Local Agent Runtime with Slack Integration and Policy Controls

David Pickett

01 Mar 2026 — 1 min read

**People:** David **Idea:** Testing OpenClaw as a self-hosted agent runtime that connects LLM agents to Slack with multi-agent routing, tool permissions, and security controls baked in. **Details:** - OpenClaw runs in a docker container and gives you channel integrations, tool execution, policy controls, and session memory out of the box - Set up a single Slack app with two routed agents - a main admin path for me and a restricted trusted path for other users - Trusted agent is locked down to workspace-scoped read operations and Slack messaging only - DM access uses an allowlist model so only specific Slack users can interact with the bot - Elevated permissions are gated to a single admin user for dangerous operations like exec - Chat commands (bash, config, restart) are disabled globally as an extra safety layer - Group channels are also on allowlist and require an @mention to trigger the bot, preventing accidental activations and bot-loop issues - Channel config works more reliably with channel IDs than human-readable channel names - Host command execution goes through an approval flow in the Gateway UI before anything runs - The architecture separates identity-based access, privilege levels, tool minimization, and execution control - which is the right pattern for multi-user Slack deployments

- Additionally we were able to get the Claude/GitHub interaction in Slack to work in tandem with the OpenClaw bot - being able to gather data from people and Claude/GitHub and send/summarize for another user

From Benchmarks to Builders: Running MiniMax M2.1 on Our Mac Studio

Last week I wrote about two paths to vLLM on Apple Silicon - comparing vllm-metal and vllm-mlx as options for local inference. This week the picture changed. LM Studio shipped concurrent request support, and suddenly the simplest option became the most practical one. People David Idea We spent the week

Connecting Attio CRM and Google Drive to Claude Desktop via MCP

People Donavan Idea Set up MCP servers for CRM and Google Drive to connect them directly to Claude Desktop - giving Claude real-time access to our deal pipeline, contacts, and documents without copy-pasting or context-switching. Details * MCP (Model Context Protocol) for connecting AI assistants to external tools and data sources.

Two paths to vLLM on Apple Silicon - vllm-metal vs vllm-mlx

People David Idea I tested both vllm-metal and vllm-mlx on the M3 Ultra Mac Studio to see which one gives better multi-user LLM throughput on a high-memory Apple Silicon machine. Details * Both projects showed up in late 2025 - vllm on Metal/MLX is very exciting because it should let

Policy-Based LLM Routing with Nvidia's Open Source Blueprint

People David Idea Testing Nvidia's v1 LLM Router blueprint as a third approach to intelligent query routing - this time using policy-based task classification instead of semantic similarity or neural network intent matching. Details * Nvidia's v1 LLM Router blueprint (main branch) takes a three-step approach: apply

Read more

From Benchmarks to Builders: Running MiniMax M2.1 on Our Mac Studio

Connecting Attio CRM and Google Drive to Claude Desktop via MCP

Two paths to vLLM on Apple Silicon - vllm-metal vs vllm-mlx

Policy-Based LLM Routing with Nvidia's Open Source Blueprint