ARCHETYPE
Minimalist
Lean instructions, few tools, tiny context tax. Lets the model think.
Copy this rig
# review before running: this installs third-party code
$ npx degit chchchadzilla/swarm-boss/kit/claude-code/.claude ./rig-swarm-boss # inspect, then merge into .claude/ MCP servers are added to Claude Code at local scope; env vars are shown as YOUR_… placeholders — we never store values. Files are fetched with degit into a separate folder so you can review before merging.
$ curl -fsSL --create-dirs -o .claude/agents/Explore.md https://raw.githubusercontent.com/chchchadzilla/swarm-boss/main/kit/claude-code/.claude/agents/Explore.md $ curl -fsSL --create-dirs -o .claude/agents/crew.md https://raw.githubusercontent.com/chchchadzilla/swarm-boss/main/kit/claude-code/.claude/agents/crew.md $ curl -fsSL --create-dirs -o .claude/agents/grunt.md https://raw.githubusercontent.com/chchchadzilla/swarm-boss/main/kit/claude-code/.claude/agents/grunt.md $ curl -fsSL --create-dirs -o .claude/agents/hater.md https://raw.githubusercontent.com/chchchadzilla/swarm-boss/main/kit/claude-code/.claude/agents/hater.md
This rig commits no guardrails. Here is the community baseline instead — the deny/ask rules most often found across all 16.2k rigs:
{
"permissions": {
"deny": [
"Read(./.env)",
"Read(./.env.*)",
"Read(**/.env)",
"Read(~/.ssh/**)",
"Bash(git push --force:*)",
"Read(**/*.pem)",
"Read(.env)",
"Read(~/.aws/**)",
"Read(**/*.key)",
"Read(**/.env.*)",
"Bash(git push --force*)",
"Bash(rm -rf:*)",
"Bash(rm -rf *)",
"Bash(sudo *)",
"Bash(sudo:*)",
"Bash(rm -rf /)",
"Bash(git push -f:*)",
"Read(.env.*)",
"Bash(git reset --hard*)",
"Bash(rm -rf /*)"
],
"ask": [
"Bash(git push:*)",
"Bash(git push *)",
"Bash(gh pr merge *)",
"Bash(git commit:*)",
"Bash(rm *)",
"Bash(gh pr merge:*)",
"Bash(gh pr create *)",
"Bash(git reset --hard:*)",
"Bash(npm publish:*)",
"Bash(git rebase *)"
]
}
} Subagents (4)
| Explore model: haiku | Fast, read-only search of files and folders -- where something lives, what a file contains, every place something is mentioned. |
| crew model: sonnet | Mid-priced worker for focused work that needs judgment inside a clear brief -- drafting a section, analyzing or clustering a chunk of findings, writing code that follows an existing pattern. Use when |
| grunt model: haiku | Low-cost worker for high-volume, mechanical work with a clear output format -- reading and tagging lots of items, extracting fields, finding every mention of something, reformatting. Use for independe |
| hater model: sonnet | Independent checker with fresh eyes. Use after workers finish and before anything gets merged or shown to the user. Finds what's wrong; doesn't fix it. |
Similar rigs
adriandlam/dotfiles
macOS dotfiles: OmniWM, Ghostty, LazyVim, Zed, zsh — linked with GNU Stow
Fort Knox 5.1k tok · · Q32
niklasrichardson/ai-tooling
Instructions, a starter kit, a skill and subagents for Claude Code, Codex, Claude and ChatGPT: short, structured, honest replies and multi-session work.
Orchestrator 4.4k tok · · Q5
Rompomepome/amiral
The admiral doesn't row. Orchestrator/worker model routing for Claude Code: a capable brain plans and verifies, cheap hands execute. Works inside your plan.
Orchestrator 1.3k tok · · Q10
ai-solution-architect-roadmap/claude-code-router
Routes each Claude Code task to the cheapest model and effort that should be enough. Rule-based and local, using the code graph, risk, past outcomes and weekly usage.
Orchestrator 349 tok · · Q15