claude-code-sub-agents
Health Gecti
- License — License: MIT
- Description — Repository has a description
- Active repo — Last push 0 days ago
- Community trust — 30 GitHub stars
Code Gecti
- Code scan — Scanned 10 files during light audit, no dangerous patterns found
Permissions Gecti
- Permissions — No dangerous permissions requested
This tool is a minimal bootstrapper for Claude Code. It installs a meta-agent, called "agent-architect," which automatically analyzes your project and generates specialized sub-agents on demand.
Security Assessment
Overall Risk: Low. The tool is lightweight and the code scan found no dangerous patterns or hardcoded secrets. It does not request overly broad or dangerous permissions. However, standard precautions apply: the primary installation method relies on piping a remote shell script directly into bash (`curl ... | bash`). While the light audit did not find malicious behavior, this execution method inherently runs arbitrary shell commands and writes local files. The script is designed to make network requests to download additional configuration files directly into your project's `.claude/` directory. No sensitive data is accessed or exfiltrated.
Quality Assessment
The project appears to be actively maintained, with recent repository pushes, and it is properly released under the permissive MIT license. It has a small but growing community trust, currently backed by 30 GitHub stars. The repository is transparent about its current scope, openly stating that it functions strictly as an installer and a single meta-agent, rather than a library of pre-built agents. It also includes continuous integration (CI) and a validation script to ensure file integrity.
Verdict
Safe to use, but as a best practice, always review the remote `init.sh` script locally before executing the bash installer.
Minimal AI agent system for Claude Code. One agent to rule them all - agent-architect creates project-specific agents on demand.
Claude Code Sub-Agents
Minimal Claude Code agent bootstrapper. It installs an agent-architect agent that can create project-specific agents inside .claude/agents/.
Current Status
This repository is a small installer plus a single meta-agent. It does not ship separate frontend, backend, DevOps, or test agent template files. Those agents are generated by agent-architect after it analyzes the target project.
Recommended profile action: keep if it reflects active Claude Code workflow work, but clean up topics. Suggested topics: claude-code, ai-agents, developer-tools, automation, shell.
Remove topics that imply implementation this repo does not contain, such as react, devops, security, and mcp.
What It Installs
Running init.sh downloads these files into the target project:
.claude/agents/agent-architect.md.claude/hooks/auto-agent-setup.sh.claude/templates/agent.template.md.claude/settings.json.claude/VERSIONCLAUDE.md
Quick Start
Review the installer before piping it to a shell:
curl -fsSL https://raw.githubusercontent.com/charles-adedotun/claude-code-sub-agents/main/init.sh
Install into the current directory:
curl -fsSL https://raw.githubusercontent.com/charles-adedotun/claude-code-sub-agents/main/init.sh | bash
Install into a specific directory:
curl -fsSL https://raw.githubusercontent.com/charles-adedotun/claude-code-sub-agents/main/init.sh | bash -s -- /path/to/project
Then ask Claude Code:
Use agent-architect to set up agents for this project
How It Works
- The installer creates
.claude/directories in the target project. - It downloads the current
agent-architect, hook, template, settings, and version files. agent-architectchecks the target project for language and framework signals.- It creates a small set of project-specific agents only when they add value.
Validation
From this repository checkout:
bash scripts/validate.sh
The validation script checks that the expected .claude files exist and that the hook is executable.
Development
bash scripts/validate.sh
License
MIT
Yorumlar (0)
Yorum birakmak icin giris yap.
Yorum birakSonuc bulunamadi