Run a personal Matrix-based assistant agent with email access
I told Kimi to set up a local Matrix server for me and to write a trivial little Claw-like.Basically it's just a loop with bubblewrap based sandboxing for the tools (not really enough if you're going to expose it to a lot of untrusted data - I intend to move the entire thing into a vm) and the "identity", "soul" and "memory" split that the original openclaw used to give the agents persistent identity.The e-mail access is read-only simply by virtue of it having no direct access to credentials. Task: I told Kimi to set up a local Matrix server for me and to write a trivial little Claw-like.Basically it's just a loop with bubblewrap based sandboxing for the tools (not really enough if you're going to expose it to a lot of untrusted data - I intend to move the entire thing into a vm) and the "identity", "soul" and "memory" split that the original openclaw used to give the agents persistent identity.The e-mail access is read-only simply by virtue of it having no direct access to credentials but access to tools to run read only operations on my mailbox.They also have tools to read and send Matrix messages, but at the end of the turn any new messages also get injected automatically if any to start a new turn. In additional there's heartbeat that triggers occasional requests to clean ensure anything they "want to remember" is consolidated in the memory files, and to clean up, or advance any tasks, and a very basic reminder system.I'm using a mix of Kimi, GLM5 to run the actual agents at the moment.I might release parts of the code at some point. It's nothing special, but it's small and easily auditable unlike OpenClaw.I picked Matrix because it's trivial to self-host with federation turned off so the agents are reasonably isolated (with the caveat that they do have web access but at least they'll have to work a bit more to leak my data - but be careful). I run it on in my wireguard network and can access it via a Matrix client on my phone and/or a web app on my home machine.I can also access it via Claude, Codex, Kimi CLI via a trivial little mcp server that wraps the same Matrix client code the agents use.; tools: OpenClaw. Human role: The original post does not say how a person stayed involved.; limitations: Prerequisites include self-hosting Matrix with federation disabled and configuring a WireGuard network. Security constraints are significant: bubblewrap sandboxing is deemed insufficient for untrusted data, requiring a VM migration. Agents retain web access, posing data leakage risks. Source reliability is secondary (Hacker News comment). Human approval is required for any action involving external communication or data modification..
Tools
Published on Agent Work Atlas:
Overview
I told Kimi to set up a local Matrix server for me and to write a trivial little Claw-like.Basically it's just a loop with bubblewrap based sandboxing for the tools (not really enough if you're going to expose it to a lot of untrusted data - I intend to move the entire thing into a vm) and the "identity", "soul" and "memory" split that the original openclaw used to give the agents persistent identity.The e-mail access is read-only simply by virtue of it having no direct access to credentials
Workflow
I told Kimi to set up a local Matrix server for me and to write a trivial little Claw-like.Basically it's just a loop with bubblewrap based sandboxing for the tools (not really enough if you're going to expose it to a lot of untrusted data - I intend to move the entire thing into a vm) and the "identity", "soul" and "memory" split that the original openclaw used to give the agents persistent identity.The e-mail access is read-only simply by virtue of it having no direct access to credentials but access to tools to run read only operations on my mailbox.They also have tools to read and send Matrix messages, but at the end of the turn any new messages also get injected automatically if any to start a new turn. In additional there's heartbeat that triggers occasional requests to clean ensure anything they "want to remember" is consolidated in the memory files, and to clean up, or advance any tasks, and a very basic reminder system.I'm using a mix of Kimi, GLM5 to run the actual agents at the moment.I might release parts of the code at some point. It's nothing special, but it's small and easily auditable unlike OpenClaw.I picked Matrix because it's trivial to self-host with federation turned off so the agents are reasonably isolated (with the caveat that they do have web access but at least they'll have to work a bit more to leak my data - but be careful). I run it on in my wireguard network and can access it via a Matrix client on my phone and/or a web app on my home machine.I can also access it via Claude, Codex, Kimi CLI via a trivial little mcp server that wraps the same Matrix client code the agents use.
Reported result
The source does not report measured outcomes, performance metrics, or specific business results. It describes the architectural intent and current setup status only. A measurable acceptance check would be verifying that the agent successfully reads emails without write permissions and maintains persistent memory across sessions.
Human role
The original post does not say how a person stayed involved.
Limitations
Prerequisites include self-hosting Matrix with federation disabled and configuring a WireGuard network. Security constraints are significant: bubblewrap sandboxing is deemed insufficient for untrusted data, requiring a VM migration. Agents retain web access, posing data leakage risks. Source reliability is secondary (Hacker News comment). Human approval is required for any action involving external communication or data modification.
Try it
Act as a DevOps engineer. I need to set up a personal AI assistant environment. 1. Deploy a local Matrix server with federation disabled. 2. Create a Python-based agent loop using bubblewrap for sandboxing tool execution. 3. Implement a file structure for 'identity', 'soul', and 'memory' to persist state. 4. Configure read-only email access tools that do not expose credentials directly. 5. Add a heartbeat script to consolidate memory and check for new Matrix messages. 6. Use Kimi or GLM5 as the LLM backend. 7. Wrap the Matrix client in an MCP server for CLI access. [Replaceable inputs: Your email provider, specific LLM API keys, target directory for code]. Before running any live email or message sending, perform a dry run with dummy data and require my explicit confirmation. Output the setup scripts and configuration files. If any security vulnerability is detected in the sandboxing, halt and report.
Source reports are not independently verified. Publication date is the catalog publication date, not the original article date.