Brendanwhit Superintendent
brendanwhit-superintendent is a Claude Code plugin for the Build phase that orchestrates autonomous Claude agents inside Docker sandboxes for clean-room, parallel agent work.
by brendanwhit · github.com/brendanwhit/superintendent
Orchestrate autonomous Claude agents in Docker sandboxes so solo builders can parallelize agent work without contaminating the main codebase.
Add it to Claude Code
Install the plugin in Claude Code. One command, paste-ready.
/plugin install brendanwhit-superintendent@brendanwhit/superintendentBuilt to be called by your agent
Skillselion is itself an MCP server. Your agent can pull this entry and a paste-ready install config straight from the API - no copy-paste.
Retrieve this entry with skillselion.get_details("plugin:brendanwhit/superintendent") and the paste-ready config with skillselion.get_install_config("plugin:brendanwhit/superintendent").
What it does
brendanwhit-superintendent is a Claude Code plugin that acts as a clean room orchestrator for spawning autonomous Claude agents inside Docker sandboxes. Solo and indie builders who already ship with Claude Code can use it when they want parallel agent work—feature branches, refactors, or experiments—without letting every agent touch the same working tree on the host. The plugin emphasizes isolation: sandboxes, orchestration, and worktree-friendly workflows so autonomous runs feel like managed workers rather than one chaotic session on your laptop. It fits the Build phase’s agent-tooling shelf but also supports safer Ship and Operate habits when you treat agent output as untrusted until reviewed. Install it when you have Docker available and you are ready to trade a bit of setup complexity for repeatable, contained agent spawning. It is not a replacement for code review or CI; it is infrastructure for running more agents at once with clearer boundaries.
Highlights
- Clean-room Docker sandboxes for autonomous Claude agent runs
- Orchestrator pattern for spawning and managing multiple agent workers
- Worktree-oriented isolation so parallel agents do not collide on one tree
- Community Claude Code plugin bundle with 1 plugin focused on sandbox spawning
- Targets builders who want agent parallelism with stronger blast-radius control than host-side runs
Why builders use it
Running multiple autonomous Claude agents on your main checkout risks mixed edits, leaked secrets, and hard-to-audit changes across one worktree.
After you register the plugin and wire Docker, you can spawn orchestrated agent runs in isolated sandboxes and promote only reviewed output back to your repo.
At a glance
- Type - Plugin in AI Agents.
- Adoption - 0 installs, 0 stars, 0 votes.
FAQ
Who is brendanwhit-superintendent for?
It is for Claude Code users who spawn autonomous agents often and want Docker-backed clean rooms instead of running everything on the host filesystem.
When should I use brendanwhit-superintendent?
Use it during Build (and when hardening Ship workflows) whenever you parallelize agent tasks and need worktree or sandbox isolation before merge.
How do I add brendanwhit-superintendent to my agent?
Install or register the brendanwhit/superintendent Claude Code plugin from the repo bundle, ensure Docker is running, then follow the plugin’s orchestration commands from your Claude Code environment.
Comments
Share how you use brendanwhit-superintendent, gotchas, or tips for other indie builders.
No comments yet - be the first to share how you use it.