Alan Fu

Building the watchdog to stop AI from misbehaving. AI is getting more permissions every day and even getting access to production code. One rm -rf and my entire drive was wiped.

Joined on Sep 3, 2026 https://www.trydoberman.dev/

Alan Fu

Sep 17

Making agent approvals easier to live with

#agents #automation #security #software

2 min read

Want to connect with Alan Fu?

Create an account to connect with Alan Fu. You can also sign in below to proceed if you already have an account.

Create Account

Already have an account? Sign in

Alan Fu

Sep 10

Fail closed: what my AI guard dog does when it doesn't know

#ai #security #opensource #python

3 min read

Alan Fu

Sep 9

Same curl, two verdicts: taint tracking for coding agents

#ai #security #devops #opensource

2 min read

Alan Fu

Sep 6

How I benchmark an agent guardrail without lying to myself

#ai #testing #security #opensource

4 min read

Alan Fu

Sep 3

Prompt filters are advisory. Enforce on the execution path.

#ai #security #llm #opensource

4 min read

Alan Fu

Sep 3

Doberman: Stops your AI before it goes rogue

#ai #security #opensource #claude

3 min read

DEV Community

Alan Fu

Badges

Writing Debut

Making agent approvals easier to live with

Want to connect with Alan Fu?

Fail closed: what my AI guard dog does when it doesn't know

Same curl, two verdicts: taint tracking for coding agents

How I benchmark an agent guardrail without lying to myself

Prompt filters are advisory. Enforce on the execution path.

Doberman: Stops your AI before it goes rogue