Claude Code
agent-cli
Anthropic's official agentic CLI. Streaming tool calls, prompt caching, thinking signatures, multi-agent subagents, slash commands.
Projects (20)
Eighteen open-source agent codebases I read in service of two of my own projects. Each profile is a normalized table — loop style, tool format, memory, sandbox — followed by drill-downs into the original analyses.
Come here when you have a codebase in mind and want the briefing. Click any concept chip to pivot into cross-project view.
Start with Claude Code — the most carefully built and best documented of the lot
OpenHands (v0)
agent-platform cloned has v2
All-hands AI v0 — autonomous software engineer agent. Event-sourced state, microagents, controller-level guardrails.
Strix
security-agent cloned
Open-source 'AI hacker' for autonomous pentesting. XML tool format, markdown-as-skills, LLM-based dedupe, module-level agent graph.
OpenHands (v1)
agent-platform cloned
OpenHands re-architected: cleaner controller, refined memory condenser, improved tool dispatch. v1 of the All-Hands agent.
Omnigent
agent-platform cloned
Open-source meta-harness: one orchestration layer over Claude Code, Codex, Cursor, Kimi, Pi and YAML-defined custom agents. Native bridges per agent, ALLOW/DENY/ASK policy gates, 8-provider cloud sandboxes, vendor-diverse code review.
Mistral Vibe
agent-cli cloned
Mistral-flavored coding agent reference. Middleware-based dispatch, minimal tool set, instructive for understanding agent loop fundamentals.
Hermes Agent
agent-framework cloned
40+ tool, multi-platform agent. Provider adapters per LLM, trajectory compression preserves first/last turns, side-channel auxiliary client.
NanoClaw
agent-cli
Tiny Claude-Code-shaped clone. Excellent for studying the irreducible structure of an agent loop without production overhead.
OpenClaw
agent-cli
Open-source Claude-Code-style agent reproduction. Bigger than NanoClaw, reveals which patterns scale and which stay minimal.
OpenKnowledge
agent-platform
Local-first WYSIWYG markdown editor that plugs INTO harnesses (Claude Code, Codex, Cursor, OpenCode) as an MCP tool-server + cross-harness SKILL.md. Agent edits are CRDT writes into the live doc; every edit is an attributed git commit on a per-writer ref. The side-channel complement to a meta-harness.
Comp AI (v2)
vertical-saas
Comp AI re-architected. Cleaner data model, refined RBAC, structured AI integrations. Useful diff target vs v1.
Claude Financial Services
vertical-saas
Reference architecture for finance-vertical Claude integrations. Patterns for compliant LLM use in regulated domains.
Comp AI (v1)
vertical-saas has v2
Compliance-as-a-service vertical SaaS. RBAC, tenant isolation, AI policy generation. v1 architecture.
Kimi Code
agent-cli cloned
Moonshot's Kimi-flavored coding agent. Compact reference for an agent loop with OpenAI-compatible tool calling.
ML Intern
agent-platform
ML-engineering-flavored agent. Tooling for data exploration, model training, and notebook-style work.
Open Design
agent-platform
Open-source design / UI-generation agent. LLM-driven design intent → code, with a design-system-aware tool surface.
Multica
agent-platform
Multi-cloud / multi-agent orchestration. Architecture patterns for spanning providers and clouds in one agent.
AIGovHub CLI
vertical-saas
AI governance CLI: detect AI usage, classify under EU AI Act, generate compliance artifacts. Vertical-saas + analyzer.
Kubiya (company analysis)
company
Not a code repo — a startup teardown. Read for product positioning, market context, and design choices in agent SaaS.
Apache Superset
data-tool
Mature open-source BI tool. Reference for data-exploration UX, plugin architecture, and large-scale Python/React structure.