AgentLoop – a Claude agent starter you can read
Readable agent loop in 150 lines, but it's a starter template competing with countless others.

Four-component breakdown demystifies agent tool-calling but ties to Go Micro framework.
Go developers building agent-enabled services
LangChain · LlamaIndex
Readable agent loop in 150 lines, but it's a starter template competing with countless others.
Enterprise security with 150+ MCP servers, but login-only landing page shows nothing.
Blog posts about building LLM agents, not a tool or framework.
Git worktree wrapper for AI agents, but worktree management is niche and solved adequately by git natively.
No-framework Python agent build that actually runs, not just theory.
Opinionated library of 16 compact skills that map real PM frameworks (Mom Test, Shape Up, Obviously Awesome, Teresa Torres) into 50–150 line agent instructions, installable via npx or a .claude plugin. Practical and immediately useful if you run AI agents for interviews, scope-cutting, or PRD drafting — the value is in the framework-to-prompt translation and the CLI/plugin delivery; it could use example outputs, tests, or usage case studies to prove the quality of each skill.