Zerg

ZERG

Zero-Effort Rapid Growth — A parallel Claude Code execution system that coordinates multiple Claude Code instances to build software features simultaneously. ZERG auto-detects your tech stack, fetches security rules, generates dev containers, breaks features into atomic tasks with exclusive file ownership, and launches a swarm of zerglings to execute them in parallel.

Why I Built This

Every time I started a new project with Claude Code, I found myself doing the same setup work over and over:

- Secure coding rules. Manually writing OWASP guidelines, language-specific security patterns, and Docker hardening rules into CLAUDE.md so Claude would actually follow them. Every. Single. Time.
- Dev containers. Configuring Dockerfiles, devcontainer.json, MCP servers, and post-create scripts so workers could run in isolated environments.
- Project scaffolding. Setting up directory structures, config files, linting, pre-commit hooks — the same boilerplate for every repo.
- Parallel execution. Claude Code is powerful, but it's one instance. For any feature with 10+ files, I'd spend hours watching a single agent work sequentially through tasks that could run in parallel.
- Context rot. The bigger the feature, the more Claude forgets. By the time it's working on file 15, it's lost track of the decisions it made in file 3.

I got tired of the repetition. So I built a system that handles all of it:

ZERG auto-detects your stack and fetches stack-specific security rules (OWASP Top 10 2025, Python, JavaScript, Docker) from TikiTribe/claude-secure-coding-rules — no manual CLAUDE.md maintenance.

ZERG generates dev containers with your detected languages, MCP server configs, and authentication baked in — workers spin up in isolated Docker environments with a single flag.

ZERG…