Every tracked project, in its proper layer.
The field guide covers the complete Agent Pulse project roster without pretending a coding product, a runtime, and an application framework are the same kind of system.
The Tracked Field
Open every dossier ›Complete scope does not mean false equivalence. The roster is divided into products, runtimes, frameworks, and specialist systems, and every page says what has and has not been inspected.
Operator-facing systems that can inspect, change, run, or review code.
4 projects Runtimes and workbenchesSystems that host, schedule, retain, or supervise agent work.
6 projects Agent frameworksLibraries for constructing agents, workflows, tools, and application-level control.
1 project Specialist agentsAgent systems whose primary execution environment is narrower than a general coding workspace.
The field guide follows every project included in Agent Pulse while keeping unlike products in distinct scope rings. A full dossier means the Atlas can teach through at least one public mechanism; it does not mean every capability has been exhaustively audited.
Coding agents
13 projectsOperator-facing systems that can inspect, change, run, or review code.
openai/codex
A local coding agent whose public Rust runtime makes session state, sandbox policy, tool modes, subagents, and operator control unusually inspectable.
google-gemini/gemini-cli
A terminal-first coding agent whose public source exposes tool scheduling, hooks, policy, context assembly, extensions, and specialized workers.
earendil-works/pi
A compact, extensible coding-agent stack whose package boundaries and session model make provider behavior, compaction, rendering, and customization legible.
charmbracelet/crush
A terminal coding agent whose public Go code exposes shell-session isolation, MCP lifecycle, configuration sharing, and the operator's live work surface.
anomalyco/opencode
An open coding agent with terminal, server, desktop, review, subagent, MCP, and provider surfaces that make the tool contract and client/runtime split especially visible.
mistralai/mistral-vibe
A terminal coding agent whose ACP compatibility, release checks, packaging, and contributor boundaries make interoperability and maturity the strongest teaching surfaces.
alejandro-ao/tau
A Python coding-agent harness whose readable architecture, extension seams, protocol changes, and contributor onboarding make the runtime itself a teaching object.
openinterpreter/openinterpreter
A Rust coding-agent fork that combines Codex lineage with local state, provider switching, and selectable harness behavior aimed at lower-cost models.
Kilo-Org/kilocode
An open coding-agent platform spanning VS Code, JetBrains, a CLI, cloud execution, code review, multiple modes, MCP, and many model providers.
QwenLM/qwen-code
A terminal coding agent that now documents IDE, desktop, daemon, SDK, messaging, multi-provider, memory, skills, sandbox, and agent-team surfaces.
cline/cline
An open coding-agent engine presented through IDE extensions, a CLI, an SDK, a multi-agent task board, scheduled work, plugins, MCP, and checkpoints.
aaif-goose/goose
A local, general-purpose agent with desktop, CLI, and API surfaces, broad provider support, and an extension model built around MCP.
OpenHands/OpenHands
A self-hosted developer control center that can run OpenHands and other ACP-compatible coding agents across local, Docker, remote, and cloud backends.
Runtimes and workbenches
4 projectsSystems that host, schedule, retain, or supervise agent work.
google-antigravity/antigravity-cli
A packaged agent CLI and migration surface whose public evidence is strongest around releases, authentication, diagnostics, permissions, and operator support rather than open runtime internals.
openclaw/openclaw
An agent runtime whose sessions, memory, command approval, scheduled wake-ups, audit trails, and channel-facing controls make long-lived operation the central teaching surface.
different-ai/openwork
An OpenCode-powered workbench that turns team memory, self-hosting, desktop packaging, authentication, MCP sessions, and shared workspaces into operator-visible product boundaries.
shepherd-agents/shepherd
A task runtime that treats an agent's output as a retained proposal, keeps authority in typed bindings, and makes settlement an explicit one-time decision.
Agent frameworks
6 projectsLibraries for constructing agents, workflows, tools, and application-level control.
langchain-ai/langchain
A broad agent application framework whose middleware, tool, model, retry, interrupt, and persistence seams expose how application code shapes an agent loop.
NousResearch/hermes-agent
A local agent framework whose approvals, worker threads, undo paths, terminal UI, telemetry, and computer-use tools expose the cost of turning broad capability into governable work.
mastra-ai/mastra
A TypeScript framework for agents and AI applications with model routing, graph workflows, suspend-and-resume control, memory, MCP, evaluation, and observability.
pydantic/pydantic-ai
A Python agent framework organized around typed dependencies, structured outputs, tools, model providers, durable execution, evaluation, and observability.
microsoft/agent-framework
A Microsoft framework for building, orchestrating, and deploying agents and multi-agent workflows across Python and .NET.
crewAIInc/crewAI
A framework centered on orchestrating role-based agents, crews, tasks, and flows for collaborative multi-agent applications.
Specialist agents
1 projectAgent systems whose primary execution environment is narrower than a general coding workspace.