Awesome Harness Engineering
Awesome list for AI agent harness engineering: tools, patterns, evals, memory, MCP, permissions, observability, and orchestration.
My AI Adoption JourneyThe essay that coined "harness engineering" (February 2026): Mitchell Hashimoto's six-stage adoption path culminates…
Harness EngineeringOpenAI's framing of harness engineering as a discipline: how to design the scaffolding that lets Codex and similar…
Unrolling the Codex Agent LoopOpenAI's detailed breakdown of the Codex agent loop, exposing each harness component and where it can be improved.
Run Long-Horizon Tasks with CodexOpenAI's practice guide for long-horizon task planning: introduces Plan.md, Implement.md, Documentation.md as reusable…
Building Effective AgentsAnthropic's foundational guide on agent architecture, covering when to use workflows vs. agents and how to compose…
Harness Design for Long-Running Application DevelopmentAnthropic's engineering blog on designing harnesses for sustained, multi-session development tasks. Key insight: every…
Writing Effective Tools for AgentsAnthropic's guide on tool interface design: naming, schemas, error surfaces, and the principle that tool design is…
Beyond Permission PromptsAnthropic on building structured permission and authorization systems into agent harnesses instead of relying on…
Demystifying Evals for AI AgentsAnthropic's framework for evaluating agent behavior: what to measure, how to build eval harnesses, and why…
What is an AI Agent?IBM's definitional piece, useful for anchoring harness design decisions to a clear model of what an agent actually is.
Agent Development Kit: Making it easy to build multi-agent applicationsGoogle's announcement and design rationale for ADK: explains the multi-agent topology, tool registration model, and…
Harness EngineeringMartin Fowler's synthesis of what harness engineering practice looks like: three interlocking systems — context…
The Anatomy of an Agent HarnessLangChain's structural breakdown of the five primitives that compose a harness: filesystem (durable state + agent…
Building AI Coding Agents for the Terminal: Scaffolding, Harness, Context Engineering, and Lessons LearnedThe first systematic practitioner paper on terminal-native coding agent harness design: eager-construction scaffolding…
Natural-Language Agent HarnessesProposes externalizing agent control logic as portable natural-language artifacts (NLAHs) executed by a shared…
Ranking Engineer Agent (REA): Meta's Autonomous AI System for Ads RankingMeta's production harness for multi-day ML pipeline automation with hibernate-and-wake checkpointing for resuming…
Supercharge Your AI Agents: The New ADK Integrations EcosystemGoogle's 2026 update to Agent Development Kit expanding the ecosystem integrations (Hugging Face, GitHub, Daytona,…
2026 Agentic Coding Trends ReportAnthropic's industry benchmark identifying infrastructure configuration as a first-class optimization variable:…
How We Build Azure SRE Agent with Agentic WorkflowsArchitecture walkthrough of Microsoft's agent that has handled 35,000+ production incidents autonomously, reducing…
Context Engineering for Reliable AI Agents: Lessons from Building Azure SRE AgentMicrosoft's account of shifting from 100+ bespoke tools and a prescriptive prompt to a filesystem-based context…
Harness Engineering: Structured Workflows for AI-Assisted DevelopmentRed Hat's enterprise perspective on harness engineering (April 7, 2026): AI writes better code when you design the…
Harness engineering for coding agent usersBirgitta Böckeler's systematic mental model (April 2026) for coding-agent harnesses, framing them as feedforward…
A Practical Guide to Building AI AgentsOpenAI's April 2026 comprehensive guide distilling production deployment patterns into actionable best practices:…
An Update on Recent Claude Code Quality ReportsAnthropic's transparent April 2026 postmortem tracing Claude Code quality degradation to three independent…
Agent Harness Design: 3 Patterns for Harnessing Claude's IntelligenceAnthropic's April 2026 design guide distilling harness engineering into three actionable patterns: build on tools…
Code as Agent HarnessMay 2026 survey framing code as the basis for agent infrastructure rather than merely output: it unifies harness…
Harness Engineering: How to Build Reliable AI Agents by Engineering the System, Not the Modeldeepset's May 2026 synthesis of agent reliability as a harness problem: a failure-classification framework (context,…
What makes a harness a harness: necessary and sufficient conditions for an agent harnessJune 2026 constitutive definition of an agent harness as a runtime layer with four necessary and sufficient elements:…
AI Harness Engineering: A Runtime Substrate for Foundation-Model Software AgentsMay 2026 formalization of the harness as the runtime substrate that mediates model–environment interaction: eleven…
Architectural Design Decisions in AI Agent HarnessesApril 2026 empirical study of 70 public agent systems across five recurring dimensions (subagent architecture, context…
RUCAIBox/awesome-agent-harnessRUCAIBox's survey paper and curated reading list on Agent Systems with Harness Engineering, mapping harness design…
Tuning the harness, not the model: a Nemotron 3 Ultra playbookLangChain's July 2026 playbook showing how harness-only tuning brought Nemotron 3 Ultra within one point of Opus 4.8…
lopopolo/harness-engineeringRyan Lopopolo's anthology, field guide, and agent context bundle for harness engineering: it reframes the harness as…
Harness Engineering for Self-ImprovementLilian Weng's July 2026 synthesis framing the harness as the deployment layer of recursive self-improvement: three…
The State Of AI Harness Engineering 2026François Zaninotto's September 2026 synthesis grounded in an audit of 246 open-source harness repos and 57…
Harness Engineering: Anatomy, Architecture, and Evolution of Coding AgentsJuly 2026 source-code study of eleven production coding harnesses (Claude Code, Codex CLI, Gemini CLI, Mistral Vibe,…
ReAct: Synergizing Reasoning and Acting in Language ModelsThe foundational paper defining the Thought/Action/Observation loop structure that underlies virtually every agent…
Unrolling the Codex Agent LoopOpenAI's detailed breakdown of the Codex agent loop, exposing each harness component and where it can be improved.
LangGraph — Low Level ConceptsModels the agent loop explicitly as a directed graph with typed state, conditional edges, and checkpointing. The most…
Unlocking the Codex Harness: How We Built the App ServerOpenAI's engineering deep-dive into the Item/Turn/Thread protocol (JSON-RPC/JSONL over stdio) that exposes the Codex…
Hooks – CodexOpenAI's lifecycle-hook framework for Codex: inject deterministic scripts at SessionStart, PreToolUse, PostToolUse,…
Extended Thinking — Claude API DocsThe harness-critical reference for integrating extended thinking into agent loops: budget_tokens controls reasoning…
Getting started with loopsAnthropic's June 2026 practical taxonomy of agent loops: turn-based, goal-based (/goal), time-based (/loop,…
Improving Deep Agents with Harness EngineeringLangChain's case study showing harness-only changes moved their coding agent from rank 30 to top 5 on Terminal Bench…
Loop EngineeringPractical design system for agent loops with seven production patterns, cross-tool starter kits, and CLI tools that…
Life-HarnessOfficial implementation of a lifecycle-aware runtime harness that improves frozen LLM agents by adapting the…
How Middleware Lets You Customize Your Agent HarnessIntroduces AgentMiddleware: six composable hooks (before_agent, before_model, wrap_model_call, wrap_tool_call,…
Agents Learn Their Runtime: Interpreter Persistence as Training-Time SemanticsControlled experiment isolating interpreter state persistence as an independent training variable. The harness…
Real-Time Deadlines Reveal Temporal Awareness Failures in LLM Strategic ReasoningDemonstrates that temporal awareness (handling deadlines and time constraints) appears orthogonal to reasoning…
A Scheduler-Theoretic Framework for LLM Agent ExecutionApril 2026 systematic analysis of 70 open-source LLM agent projects showing 60% adopt the Agent Loop pattern. Proposes…
Confucius Code Agent (CCA)February 2026 production-grade coding agent from Meta/Harvard built on the Confucius SDK, which structures harness…
The Design Space of Today's and Future AI Agent SystemsApril 2026 reverse-engineering of Claude Code's architecture revealing five-stage progressive compaction (budget…
deepclaudePorts Claude Code's full agent loop to DeepSeek V4 Pro and other Anthropic-compatible backends while preserving the…
The Coding Harness Behind GitHub Copilot in VS CodeVS Code team's breakdown of the coding harness behind GitHub Copilot: three core loop responsibilities (context…
statewrightState machine guardrails that constrain which tools an agent can call in each phase of a workflow, turning open-ended…
Introducing dynamic workflows in Claude CodeAnthropic's May 2026 introduction to dynamic parallel subagent orchestration: Claude generates JavaScript…
AgentSPEXUIUC's open-source specification and execution language for LLM-agent workflows: declarative YAML with typed steps,…
TomasuLLM: Out-of-Order Speculative Execution for LLM AgentsSeptember 2026 paper that ports speculative execution from processor design to the agent loop: the harness predicts…
Run Long-Horizon Tasks with CodexOpenAI's practice guide for long-horizon task planning: introduces Plan.md, Implement.md, Documentation.md as reusable…
Harness Design for Long-Running Application DevelopmentAnthropic's engineering blog on designing harnesses for sustained, multi-session development tasks. Key insight: every…
Plan-and-Execute AgentsThe canonical engineering write-up separating planning from execution as distinct harness layers: a planner LLM…
microsoft/TaskWeaverCode-first task decomposition framework with a planner/executor split and a plugin system for injecting domain…
LATS: Language Agent Tree SearchUnifies reasoning, acting, and planning via Monte Carlo Tree Search over agent trajectories. Directly informs harness…
Agyn: A Multi-Agent System for Team-Based Autonomous Software EngineeringDemonstrates specialized harness patterns for coordinating heterogeneous agent teams (planner, coder, reviewer,…
Plan-and-Act: Improving Planning of Agents for Long-Horizon TasksModular framework separating high-level planning from low-level execution through synthetic data generation and…
Choosing the Right Multi-Agent ArchitectureDecision framework for four multi-agent patterns (subagents, skills, handoffs, router) with concrete performance data:…
Multi-Agent Workflows Often Fail. Here's How to Engineer Ones That Don't.GitHub's February 24, 2026 distillation of a failure pattern most harnesses eventually rediscover: multi-agent systems…
Effective Harnesses for Long-Running AgentsAnthropic's pattern for maintaining agent progress across multiple context windows: an initializer agent sets up the…
Task-Adaptive Multi-Agent Orchestration (AdaptOrch)February 2026 framework that dynamically selects orchestration topology (parallel, sequential, hierarchical, or…
GitHub Spec KitGitHub's open-source toolkit for spec-driven development: spec.md → plan.md → tasks.md artifacts convert…
Task-Decoupled Planning for Long-Horizon Agents (TDP)January 2026 planning framework that combines task decomposition with modular agent design: a Supervisor decomposes…
Fresh Memory, Stale Plans: Derivation Currency for Distributed LLM-Agent MemorySeptember 2026 paper naming the failure mode every freshness check misses: an agent can inherit the latest requirement…
Harness EngineeringOpenAI's framing of harness engineering as a discipline: how to design the scaffolding that lets Codex and similar…
Effective Context Engineering for AI AgentsAnthropic's systematic guide to managing the full context state—system prompts, tools, MCP, and message history—as a…
Compaction — Claude API DocsAnthropic's reference for server-side context compaction: automatically summarizes older context when approaching the…
LLMLinguaMicrosoft Research's prompt compression toolkit (up to 20x compression, minimal performance loss) that can be embedded…
Prompt Caching — Claude API DocsThe most effective harness-level cost lever: cache repeated system prompts, tool definitions, and long documents…
Autonomous Context CompressionShifts context compression from harness-controlled (compacting at a fixed token threshold) to agent-controlled: agents…
Active Context Compression: Autonomous Memory Management in LLM AgentsProposes a "Focus Agent" architecture where the agent autonomously decides when to consolidate interaction history…
context-modeMCP server that intercepts raw tool output before it enters the context window, sandboxing bulky data (Playwright…
Making Agent-Friendly Pages with Content NegotiationVercel's February 3, 2026 implementation guide for serving text/markdown when agents request it via Accept:…
A-RAG: Scaling Agentic Retrieval-Augmented Generation via Hierarchical Retrieval InterfacesReframes RAG as a harness tool-design problem: instead of injecting retrieved documents into context at pipeline time,…
LLM Readiness Harness: Evaluation, Observability, and CI Gates for LLM/RAG ApplicationsStructured framework for building production-grade evaluation harnesses: evaluation gates that block deployment,…
ByteRover: Agent-Native Memory Through LLM-Curated Hierarchical ContextLLM-curated hierarchical context management for agents where the model itself learns to weight information importance…
Claude Code Compaction: How Context Compression WorksMarch 2026 deep-dive into Claude Code's automatic compaction mechanism: what survives (current task, recent errors,…
Token SaviorMCP server that indexes codebases by symbol (functions, classes, call graphs) so agents navigate by pointer instead of…
TrellisReplaces the bloated CLAUDE.md pattern with a progressive spec system: agents load only the standards, task PRDs, and…
OpenVikingByteDance/Volcengine's open-source context database that puts an agent's knowledge, memory, and skills in one virtual…
DESIGN.mdGoogle Labs' specification for describing visual identity systems to coding agents: machine-readable design tokens…
codebase-memory-mcpHigh-performance code intelligence MCP server that full-indexes repositories into a persistent knowledge graph via…
MirageMounts S3, Slack, Gmail, GitHub, and Redis side-by-side as a single virtual filesystem so agents interact with every…
diracCoding agent harness optimized for surgical context curation and API cost reduction: Hash Anchored edits, massively…
MinishLab/sembleCode search primitive that replaces grep+read cycles with natural-language retrieval, cutting active tokens by ~98%…
harness-experimentalRepository-level operating harness that turns any software repo into an agent-ready workspace: structured AGENTS.md,…
headroomCompresses tool outputs, logs, files, and RAG chunks before they enter the context window, cutting active tokens by…
Context7MCP server and CLI that injects up-to-date, version-specific library documentation directly into agent context,…
Context Pruning for Coding Agents via Multi-Rubric Latent ReasoningDecomposes code relevance into two interpretable dimensions — semantic evidence and dependency support — rather than…
OpenWikiLangChain's CLI that writes and maintains agent-readable wikis for codebases or purpose memory, turning documentation…
PRO-LONGProgrammatic memory framework for long-horizon agents: the harness appends all observations to a structured log and…
GraftBuilds a local, regenerable graph of plain-English system explanations and code relationships, then rides along inside…
ktxSelf-improving executable context layer for data and analytics agents: it ingests warehouses, BI tools, and wikis to…
Organizing Context in a Multi-Agent HarnessLangChain's September 2026 introduction of context modes for subagents: isolated subagents start fresh (context…
SoL-PiNVIDIA's September 2026 extension for the Pi harness that packages four efficiency mechanisms discovered by running…
Boris Cherny: Building Claude Code (YC Startup School, July 2026)The Claude Code creator's account of shipping Opus 5: Anthropic deleted ~80% of the system prompt because it encoded…
Context management in agent harnesses: memory, files, and subagentsArize's April 2026 comparative teardown of how Pi, OpenClaw, Claude Code, and Letta each manage context as an active…
Context Engineering in 2026: Why We Stopped Compacting Our Agent's ContextAugust 2026 controlled study on a production AI tutor (660 turns, planted-fact probes) that inverts the compaction…
AutoCompact: Learning When to Compact Context in Long-Horizon Coding AgentsOctober 2026 paper that turns compaction timing from a fixed harness policy into a learned agent behavior: a judge…
Context Compaction TheoryAugust 2026 paper (Tirmazi, Markelon, Bishop, Mitzenmacher) that initiates the formal study of compaction: the minimum…
The Compaction Cliff in Long-Running AI Agent MemoryCIKM 2026 paper measuring what compaction actually destroys: Claude Code's /compact prompt on Sonnet 4.6 retains only…
Diagnosing and Mitigating Context Rot in Long-horizon SearchGAIR/SJTU paper that names the failure mode compaction debates usually miss: premature termination, where search…
AGENTS.md vs CLAUDE.md: which file each harness readsSide-by-side of how Claude Code, Codex, OpenCode (v1 and v2), goose, Aider and Gemini CLI discover and prioritize repo…
Slipstream: Trajectory-Grounded Compaction Validation for Long-Horizon AgentsMay 2026 paper attacking compaction's structural validation gap: synchronous compaction must condense context without…
A failed network check is a finding, not a tool crashA worked Python adapter separates a diagnostic finding from an execution error when a CLI returns valid JSON with exit…
Writing Effective Tools for AgentsAnthropic's guide on tool interface design: naming, schemas, error surfaces, and the principle that tool design is…
Tool Use — Claude API DocsAuthoritative reference for client vs. server tool execution models, strict schema enforcement, and tool_result error…
Function Calling — OpenAI DocsDefines the de facto industry-standard JSON Schema conventions for tool definitions and parallel function calling.…
Tool Annotations as Risk VocabularyThe MCP team's definitive post on the four tool annotation hints (readOnlyHint, destructiveHint, idempotentHint,…
outlinesConstrains token sampling via regex/CFG/JSON Schema at the decoding layer, guaranteeing structured output without…
instructorMaps Pydantic models directly to structured LLM extraction with built-in retry and validation-error feedback loops.…
SkillTester: Benchmarking Utility and Security of Agent SkillsFramework for evaluating agent skills on three dimensions (capability, robustness, security) before deployment.…
AutoHarness: Improving LLM Agents by Automatically Synthesizing a Code HarnessGoogle DeepMind technique that uses code synthesis to auto-generate runtime constraint harnesses from tool schemas and…
Scaling Parallel Tool Calling for Efficient Deep ResearchFebruary 2026 analysis of how parallel tool calling reduces latency in multi-step agent workflows. Demonstrates that…
EigentSearch-Q+April 2026 framework for deep-research agents using dedicated reasoning tools (plan_next_searches,…
TopoCurate: Modeling Interaction Topology for Tool-Use Agent TrainingMarch 2026 framework that models interaction topology — the structural patterns of how agents invoke, chain, and…
Design Patterns for Deploying AI Agents with Model Context ProtocolMarch 2026 field report from an enterprise MCP deployment identifying three protocol-level gaps that break production:…
tui-useExpands the agent tool surface beyond non-interactive commands: programmable TUI interaction for REPLs, debuggers, and…
CLI-AnythingGenerates agent-native CLI harnesses for any software, giving agents structured JSON access to applications that were…
zerolangExperimental graph-first programming language where agents inspect and edit code through a compiler-derived…
The Devil Is in the Interface: Evaluating How Tool Architecture Shapes Coding Agent BehaviorAugust 2026 controlled study (11,700 trajectories, six tool architectures with capabilities held constant) isolating…
Model Context ProtocolAnthropic's open protocol for connecting agents to external tools, data sources, and services in a standardized way.
modelcontextprotocol/serversAnthropic's official reference MCP server implementations (GitHub, Slack, Postgres, Puppeteer, etc.). The…
microsoft/playwright-mcpBrowser automation via accessibility tree snapshots rather than screenshots, dramatically reducing token cost. The…
Chrome DevTools MCPOfficial Google MCP server that exposes live Chrome debugging surfaces — network analysis, performance profiling,…
agent-deviceMCP-native control layer for iOS and Android devices: snapshots, semantic targeting, typed client access, diagnostics,…
A2A ProtocolGoogle's open Agent-to-Agent protocol: JSON-RPC over HTTP(S)/SSE with Agent Card service discovery and a…
Announcing the Agentic Resource Discovery specificationGoogle's June 2026 open specification for publishing, discovering, and verifying AI capabilities across the web via…
MCP InspectorInteractive debugging UI for MCP servers: inspect tool definitions, send test calls, and validate responses without…
Shell + Skills + Compaction: Tips for Long-Running AgentsOpenAI's engineering guide to three production harness primitives: versioned Skill bundles (SKILL.md manifest; routing…
ComposioWraps 250+ SaaS APIs (GitHub, Slack, Linear, Notion, etc.) as agent-ready actions with managed OAuth, so tool…
oomol-lab/open-connectorOpen-source auth gateway connecting 1000+ SaaS providers to AI agents through SDK, CLI, MCP, HTTP, and OpenAPI. Treats…
MCP Streamable HTTP TransportThe transport that replaced HTTP+SSE in the 2025-11-25 spec, enabling MCP servers to run as remote services rather…
The 2026 MCP RoadmapThe MCP team's roadmap for the next spec cycle: horizontal-scaling transport without stateful session constraints,…
The 2026-07-28 MCP Specification Release CandidateThe largest revision of MCP since launch: a stateless protocol core drops the initialize handshake and Mcp-Session-Id,…
Developer's Guide to AI Agent ProtocolsGoogle's survey of six standardized agent interoperability protocols, each solving a distinct harness integration…
AG-UILightweight event-driven protocol standardizing how AI agents connect to frontend applications: streaming state…
Agent Client Protocol (ACP)Zed and JetBrains' open standard (Apache-2.0) applying the LSP model to coding agents: JSON-RPC 2.0 over stdio, where…
Code Execution with MCP: Building More Efficient AgentsAnthropic's engineering account of reducing tool-call token overhead by having agents write code to interact with MCP…
Microsoft Skills FrameworkStandardized framework for defining, versioning, and distributing agent skills. Enables skill reuse across Claude…
SkillNet & SkillsBench: Infrastructure for AI Agent Skills at ScaleComprehensive framework for creating, evaluating, and sharing agent skills with 86-task benchmark across 11 domains.…
CometResumable long-running task workflow and Skill platform for coding that turns skill creation, evaluation, and release…
addyosmani/agent-skillsProduction-grade engineering skills for AI coding agents, packaged as 24 reusable skills covering the full development…
AWS Bedrock AgentCore with WebRTC SupportAdds peer-to-peer, UDP-based WebRTC bidirectional streaming to Bedrock Agents for real-time voice interactions.…
Hermes Agent: Unified Streaming for Real-Time Agent WorkflowsToken-by-token streaming delivery system enabling real-time agent responses; sub-second decision loops on streaming…
Google Developers: Closing the Knowledge Gap with Agent SkillsGoogle ADK expansion with evaluation harness (117 prompts) for assessing skill performance across agentic coding,…
Agent Skills Can Be Harmful: An Empirical Study of Skill-Induced Failures in LLM AgentsAugust 2026 study (authors including Tianyin Xu, UIUC) that attributes 307 skill-induced failures from SkillsBench and…
What's New with GitHub Copilot Coding AgentGitHub's February 26, 2026 update is worth including for one specific reason: it makes .github/agents/ custom agent…
Announcing Official MCP Support for Google ServicesGoogle's 2026 rollout of managed MCP endpoints is a useful counterpoint to self-hosted MCP servers: discovery, IAM,…
Dataverse Skills: Your Coding Agent Now Speaks DataverseMicrosoft's April 1, 2026 release is a strong concrete example of domain-specific skills done properly: the agent…
You can't whisper at an AI agentStripe's May 2026 study of how agents actually consume SDK and CLI guidance: passive documentation is ignored, while…
Agent Toolkit for AWSOfficial AWS-supported MCP servers, skills, and plugins that let AI agents provision, query, and manage AWS resources…
agentic-stackPortable .agent/ folder that externalizes memory, skills, and protocols from any specific coding agent into a…
mcp-agentProduction-grade framework for building agents with MCP: composable workflows, built-in observability, and…
vurb.tsTypeScript framework for building production MCP servers with a "Presenter" perception layer that strips undeclared…
SkillOptMicrosoft's skill optimizer that trains reusable natural-language skills for frozen LLM agents through…
superpowersAgentic skills framework and software-development methodology whose session-start hook loads a meta-skill that makes…
Antigravity Awesome SkillsInstallable library of 1,400+ agentic skills for Claude Code, Cursor, Codex CLI, Gemini CLI, and more. The largest…
agentgatewayOpen-source agentic proxy that unifies LLM gateway, MCP gateway, and A2A gateway into a single control plane for…
AIP: A Graph Representation for Learning and Governing Agent SkillsJune 2026 proposal to replace free-form skill prose with directed execution graphs: discrete steps as nodes backed by…
PonytailSkill that makes coding agents behave like a "lazy senior dev": prefer built-in solutions, avoid new dependencies, and…
wshobson/agentsCross-harness plugin marketplace that maintains one source-of-truth plugins/ directory and generates harness-native…
mgechev/skillgradeCLI that turns agent skill verification into repeatable unit tests: it generates task and grader pairs from a…
Qwen-MM-PluginsQwen's official multimodal plugin suite packages vision, video, document, 3D, and CAD capabilities as portable skills…
Agent-Loop-SkillsTurns the agent loop itself into portable Agent Skills: 25 verification-gated iteration loops (autoresearch,…
Mods overview — Claude Code DocsAnthropic's October 2026 extension primitive that moves harness customization from outside the loop to inside it: a…
Claude Design Skills TestedAgentCraft's same-brief test of the 8 most installed Claude design skills plus a no-skill control (Claude Haiku, a…
Oh My AndroidNative macOS control panel for the Android Emulator (dark mode, font scale, locale, TalkBack, network throttling, GPS…
FirecrawlMCP server for web search, scraping, and browser interaction that returns pages as Markdown or as JSON matching a…
MCP servers over stdio: startup time, call latency and failuresAgentic Weekly timed 15 MCP server setups over stdio with the official Python SDK (cold start, tools/list, median and…
Beyond Permission PromptsAnthropic on building structured permission and authorization systems into agent harnesses instead of relying on…
OWASP LLM06:2025 — Excessive AgencyOWASP's authoritative definition of the "excessive agency" risk: over-provisioned functions, unnecessary permissions,…
GitHub Enterprise — Governing AgentsApril 2026 GitHub official guide for enterprise agent governance: MCP server registry curation with ruleset-protected…
Claude Code Auto Mode: A Safer Way to Skip PermissionsAnthropic's engineering post on replacing approval fatigue (users approve 93% of prompts, making approvals…
Claude Agent SDK — Configure PermissionsThe most concrete reference for harness permission architecture: five-layer evaluation order (hooks → deny rules →…
Two Different Types of Agent AuthorizationDistinguishes on-behalf-of authorization (agent uses end-user credentials, requires cross-channel identity mapping and…
Authorization and Governance for AI Agents: Runtime Authorization Beyond Identity at ScaleMicrosoft Security's reusable Authorization Fabric combining a Policy Enforcement Point (PEP) and Policy Decision…
IETF draft-klrc-aiagent-auth: AI Agent Authentication and AuthorizationThe first IETF standards-track specification for AI agent authentication (March 2026, authors from AWS, OpenAI,…
Nango: Pre-Built Authentication for AI AgentsOpen-source platform providing pre-built OAuth and API key authentication for 700+ APIs across 30 categories.…
Agent VaultInfisical's open-source credential broker that sits between AI agents and the APIs they call, injecting real…
AgentDoG: A Diagnostic Guardrail Framework for AI Agent Safety and SecurityThree-dimensional risk taxonomy (source/failure-mode/consequence) with fine-grained agentic safety benchmark (ATBench)…
Open Agent Passport (OAP): Deterministic Pre-Action Authorization for Autonomous AI AgentsMarch 2026 open specification and reference implementation that intercepts tool calls synchronously before execution,…
nahDeterministic permission guard that maps tool calls to an intent taxonomy (filesystem_delete, network_outbound,…
AegisPre-execution firewall that intercepts, classifies, and blocks agent tool calls before they execute, with a compliance…
When "Do Not" Is Not Deny: Security Rules in CLAUDE.md vs Built-In ControlsAnalyzes 481 public CLAUDE.md files and finds only ~4% of natural-language security rules are backed by a matching…
Do User-Authored Permission Policies Improve Protection Against AI Agent Overreach?August 2026 user study (113 non-technical participants) comparing per-action HITL approval, automated model review,…
Connections: Managed Credentials and Per-Caller Identity for Managed Deep AgentsLangChain's September 2026 implementation of on-behalf-of authorization as a harness primitive: credentials live in…
Building Effective AgentsAnthropic's foundational guide on agent architecture, covering when to use workflows vs. agents and how to compose…
Letta (MemGPT)The reference architecture for stateful agents: three-tier memory (core / archival / recall) maps directly to harness…
mem0Drop-in universal memory layer (YC-backed, AWS Agent SDK's exclusive memory provider) that handles cross-session…
StashSelf-hosted persistent memory layer with an 8-stage consolidation pipeline (episodes → facts → relationships →…
TencentDB-Agent-MemoryTencent's fully local agent memory system with a 4-tier progressive pipeline (Conversation → Atom → Scenario →…
ZepPurpose-built agent memory store with automatic conversation summarization, entity extraction, and semantic search…
engramPersistent memory for AI coding agents delivered as a single Go binary with SQLite + FTS5 and 18 MCP tools for save,…
MemPalaceLocal-first AI memory system that stores conversation history verbatim and retrieves it with semantic search through a…
agentmemoryPersistent memory server for coding agents that retrieves only the top-K relevant observations via hybrid BM25 +…
claude-memory-compilerTurns raw Claude Code sessions into a self-evolving knowledge base: hooks capture every interaction, the Agent SDK…
mexTurns agent-learned project knowledge into a repo-local, symbol-grounded wiki with drift detection and task-aware…
How We Built Agent Builder's Memory SystemLangChain's engineering account of a COALA-based three-tier memory system (procedural/semantic/episodic) backed by…
When Malicious Instructions Persist: Persistent Memory Poisoning Attack on Harness-Based AgentsSeptember 2026 attack paper demonstrating that untrusted content tricked into a coding agent's persistent memory…
Building an Agentic Memory System for GitHub CopilotGitHub's January 15, 2026 write-up is one of the clearest public discussions of deployed cross-agent memory:…
Building Self-Correcting Memory in OpenWikiLangChain's August 2026 design account of memory that can forget: every claim the agent writes is persisted with…
MemArchitect: A Policy-Driven Memory Governance LayerProposes a governance layer that decouples memory lifecycle management (decay, conflict resolution, privacy…
Codified Context: Infrastructure for AI Agents in a Complex CodebaseProduction-validated architecture (283 sessions, 108k-line codebase) built on three components: a "hot-memory…
Facts as First Class Objects: Knowledge Objects for Persistent LLM MemoryIdentifies three production failure modes of in-context memory at scale: capacity overflow at ~8,000 facts, 60% fact…
Governance Decay: How Context Compaction Silently Erases Safety Constraints in Long-Horizon LLM AgentsJune 2026 paper establishing context compaction as a first-class governance surface: the ConstraintRot benchmark…
Recoverability Has a Law: The ERR Measure for Tool-Augmented AgentsFormal framework for measuring how well agents recover from tool failures. Defines Expected Recovery Regret (ERR) as a…
MAGMA: Multi-Graph Agentic Memory ArchitectureRepresents agent memory across four orthogonal semantic, temporal, causal, and entity graphs, enabling policy-guided…
GAAMA: Graph Augmented Associative Memory for AgentsHybrid memory system blending graph traversal with semantic similarity through additive scoring; graph augmentation…
Graph-Native Cognitive Memory for AI Agents: Formal Belief Revision Semantics for Versioned Memory ArchitecturesFormal semantics for versioned memory graphs with belief revision operations, enabling agents to maintain coherent…
Continual learning for AI agentsLangChain's April 2026 framing of agent learning as three distinct layers: model weights, harness behavior, and…
cogneeOpen-source memory platform with a hybrid graph-vector-relational poly-store that lets agents recall facts through…
HindsightAgent memory system organized around three explicit operations—retain, recall, and reflect—with semantic, keyword,…
ClawVM: Harness-Managed Virtual Memory for Stateful Tool-Using LLM AgentsApplies virtual-memory semantics to agent context management, treating the context window as working memory with typed…
MAGE: Memory as Agent-Guided ExplorationJune 2026 proposal to treat long-horizon memory as execution-state management rather than semantic retrieval: a…
deja-vuIndexes coding-agent sessions already written to disk and serves them back over MCP with no LLM calls, embeddings, or…
trajectoryLetta's July 2026 library that normalizes native session transcripts from 15+ harnesses (Claude Code, Codex, Cursor,…
Harness the Memory: A Holistic Evaluation of Memory Substrates in Memory AgentsAugust 2026 controlled study running every common memory substrate (dense/sparse indices, text records, structural and…
Who Remembers What? Multi-Granularity Memory Extraction for Multi-User, Multi-Agent InteractionNovember 2026 paper (MixUA) on memory writes in harnesses shared by many users and heterogeneous agents: each lesson…
MnemoverseMIT-licensed MCP server for agent memory, backed by a hosted engine. For harness design it shows the recall loop…
Senrelay (ContextCore)Local-first macOS harness workbench combining raw cross-agent session relay (/contextcore bounded transcript paging…
LightDAG (LightAgent)Single-host dynamic task graphs with SQLite persistence, bounded concurrent workers, and restart recovery through…
Harness EngineeringOpenAI's framing of harness engineering as a discipline: how to design the scaffolding that lets Codex and similar…
Building a C Compiler with a Team of Parallel ClaudesAnthropic's account of coordinating 16 Claude instances in parallel on a shared git repo without a central…
LiteLLMUnified proxy and SDK that routes to 100+ LLM providers behind a single OpenAI-compatible interface, with a Router…
LangGraphGraph-based state machine framework for multi-agent harnesses: models supervisor/subagent topologies, error-recovery…
OpenAI Agents SDKLightweight multi-agent framework built around handoffs and guardrails; the production successor to Swarm. Complements…
Codex SDKOpenAI's official SDK for programmatically controlling local Codex agents from TypeScript or Python: start threads,…
Codex as a platform: build on the open agent harnessOpenAI's September 2026 guide to treating the open-source Codex harness as reusable infrastructure rather than a…
Google ADKGoogle's code-first agent framework with built-in multi-agent orchestration, tool registration, session state, and…
strands-agents/harness-sdkAWS's open-source, model-driven agent SDK that treats the loop, tool binding, and guardrails as first-class…
Build Long-running AI agents that pause, resume, and never lose context with ADKGoogle's May 2026 guide to production agents that survive days-long idle periods: DatabaseSessionService for…
Build Cross-Language Multi-Agent Team with Google's Agent Development Kit and A2AGoogle's June 2026 production pattern decomposing monolithic agents into cross-language specialists: a Python LLM…
AutoGenMicrosoft's multi-agent conversation framework with a complete AgentChat layer covering agent loop, tool integration,…
CrewAIDual-layer harness orchestration: Crew handles autonomous agent delegation, Flow provides event-driven deterministic…
Pydantic AI v2June 2026 harness-first redesign built around the Capability primitive: a single composable unit bundling…
OmniRoute: Multi-Provider LLM GatewayIntelligent routing across multiple LLM providers with load balancing, intelligent fallbacks, rate limiting, and…
OpenSquillaToken-efficient microkernel agent with an on-device SquillaRouter that dispatches each turn to the cheapest capable…
Scaling Managed Agents: Decoupling the Brain from the HandsAnthropic's production architecture for separating three stateless components — the "brain" (Claude + harness),…
Microsoft Agent Framework at BUILD 2026: Agent Harness, Hosted Agents, CodeAct, and moreMicrosoft's June 2026 expansion folds previously scattered harness primitives into a single runtime layer: automatic…
Microsoft Agent Framework 1.0Production-ready 1.0 release (April 2026) unifying Semantic Kernel and AutoGen into a single framework with…
ConductorMicrosoft’s open-source YAML-first CLI for deterministic multi-agent orchestration: Jinja2 routing, static parallel…
Shopify RoastShopify's open-source Ruby DSL for structured AI workflows that interleaves deterministic steps (shell commands, Ruby…
AgentScope RuntimeProduction-ready open-source runtime focused on two pieces many agent frameworks leave underspecified: secure sandbox…
AgentScope JavaProduction-ready Java framework for distributed, enterprise-grade agents with workspace sandboxing, permission-gated…
Orchestrating Ambient Agents with TemporalTemporal.io's harness infrastructure for persistent agent workflows with native agentic handshake protocol for secure…
Vercel AI SDKThe leading TypeScript toolkit for building AI agents (20M+ monthly downloads, 25+ provider integrations). AI SDK 6…
eveVercel's filesystem-first framework for durable AI agents: instructions, typed tools, on-demand skills, message…
MastraTypeScript-native agent framework (from the Gatsby team) with 22K+ stars and 300K+ weekly npm downloads. Connects to…
FlueAstro's TypeScript-native agent harness that treats an agent as a function composed of model, sandbox, skills, tools,…
open-multi-agentTypeScript-native multi-agent orchestration that automatically decomposes a natural-language goal into a task DAG,…
The next evolution of the Agents SDKOpenAI's April 2026 update adding native sandbox execution, configurable memory, and sandbox-aware orchestration to…
SymphonyOpenAI's orchestration layer for moving from supervising coding agents to managing work: it monitors an issue tracker,…
HarmonistDrop-in multi-agent framework where protocol enforcement is a mechanical gate, not a prompt request: IDE-level hooks…
HiveYC-backed production harness that compiles multi-agent objectives into deterministic execution DAGs with state…
thClawsNative-Rust agent harness platform with four surfaces (GUI, CLI, web, one-shot), multi-provider routing, and…
sandcastleTypeScript-native orchestration for sandboxed coding agents that treats provider-agnostic isolation (Docker, Podman,…
bernsteinDeterministic scheduler for 40+ CLI coding agents running in parallel git worktrees with an HMAC-signed audit chain,…
waggleAttributed, resolvable artifact references for agent handoffs: a ~30-byte token replaces pasted context, letting each…
OrcaThe AI orchestrator for running Codex, Claude Code, OpenCode, and Pi side-by-side in isolated git worktrees. It turns…
LoopXProvider-neutral control plane that sits above existing harnesses (Codex, Claude Code, Cursor) and keeps long-horizon…
HarnessRouterSelf-hosted unified API for running Codex, Claude Code, DeepSeek Harness, Pi, and other terminal agents through the…
ScionGoogle's experimental "hypervisor for agents" (April 2026, Apache-2.0): orchestrates Claude Code, Gemini CLI, Codex,…
Sno StationLocal runtime with no daemon that makes Codex, Claude Code, Cursor, Hermes Agent and OpenClaw work as one team: sno…
OrchestratorOpen-source desktop app (macOS, Windows, Linux, with iPhone and Android companions) where an orchestrator agent splits…
KraftLocal orchestrator that keeps the process outside the agent: a YAML chain file sets the order of agent, command and…
Atomic Agent FusionOrchestrator and worker split inside an MIT local-first agent: the orchestrator only plans, writes a brief per part…
Demystifying Evals for AI AgentsAnthropic's framework for evaluating agent behavior: what to measure, how to build eval harnesses, and why…
promptfooYAML-driven LLM testing framework with LLM-as-judge, assertion DSL, and native CI integration. The most practical tool…
AgentBenchMulti-environment agent benchmark (OS, DB, web, code) with a structured eval pipeline. Worth studying for its…
Testing Agent Skills Systematically with EvalsOpenAI's framework for skill regression testing: four eval dimensions (outcome, process, style, efficiency goals),…
Agent Evaluation Readiness ChecklistA 33-item checklist covering the full evaluation lifecycle: error taxonomy, three-level granularity (single-step →…
Evaluating SkillsLangChain's methodology for benchmarking agent skills in Docker-sandboxed environments. Key empirical findings: Claude…
AgentAssay: Token-Efficient Regression Testing for Non-Deterministic Agent WorkflowsAddresses agent CI's core problem: binary pass/fail is useless for non-deterministic workflows. Behavioral…
Agentic Harness for Real-World Compilers: A Case Study in Specialized Tool DesignDemonstrates how harness specialization (llvm-autofix) for a narrow domain (compiler bug fixes) achieves better…
Eval-Driven Development: Build and Evaluate Reliable AI AgentsRed Hat's eight-stage evaluation maturity progression from manual CLI testing to cost-aware continuous monitoring…
Agent Evaluation Framework 2026: Metrics, Rubrics & BenchmarksComprehensive framework combining multi-environment baselines (AgentBench), domain-specific benchmarks (Terminal Bench…
The 2025 AI Agent Index: Documenting Technical and Safety Features of Deployed Agentic AI SystemsSystematic index analyzing deployed agent safety documentation, guardrails, and third-party testing across 30…
sentruxReal-time architectural sensor that closes the feedback loop for coding agents: scans codebases, scores structural…
Driving the Agent Quality Flywheel from Your Coding AgentGoogle's June 2026 account of automating the eval-optimize loop for coding agents: independent AutoRaters grade agent…
mcp-test-harnessPytest-style testing framework for MCP servers with protocol-aware assertions, snapshot tests, and CI-ready reports…
shadcn-ui/lintSeptember 2026 agent-first linter for Tailwind design systems: you encode design-system contracts as programmable lint…
SLCR (Sliceable, Linked, Checkable Requirements)An open requirements format and CLI that maps each requirement’s acceptance criteria to validators (tests, linters,…
AEGISWindows-first desktop monitor for local agent processes, configured file activity, TCP endpoints and audit exports.…
OpenLLMetryOpenTelemetry-based instrumentation for LLM calls and agent steps: adds trace spans to every inference and tool call…
Arize PhoenixSelf-hostable trace UI and eval runtime for agent workflows. Lets harness engineers audit and replay every reasoning…
OpikComet's open-source AI observability and evaluation platform: deep tracing of LLM calls, conversation logging, and…
LangfuseThe most widely adopted self-hostable LLM observability platform: traces every agent step, manages prompt versions,…
Weights & Biases WeaveW&B's tracing and eval layer purpose-built for agent workflows: automatic call graph capture, dataset versioning, and…
OTel GenAI Semantic ConventionsOpenTelemetry's standard attribute names for GenAI spans (gen_ai.system, gen_ai.request.model, etc.). The naming…
Pydantic LogfireAI observability platform from the Pydantic team with a unique angle: all trace data is SQL-queryable…
HeliconeOpen-source LLM observability proxy (YC W23) with the largest open-source pricing database (300+ models). One-line…
OpenObserve: Unified Observability for LLM Agents2026-standard platform for LLM tracing with infrastructure log/metric unification. Enables harness engineers to…
BraintrustEvaluation-first agent observability platform ($80M Series B, Feb 2026) with exhaustive auto-tracing that captures…
Building Observable AI Agents: Temporal Now Integrates with BraintrustCombines Temporal's durable execution (automatic retries, state persistence, event history replay) with Braintrust's…
Introducing BigQuery Agent AnalyticsGoogle Cloud's 2026 launch treats agent traces, tool calls, sessions, and outcomes as analytical data rather than…
Distributed Tracing for Agentic Workflows with OpenTelemetryRed Hat's April 6, 2026 guide is one of the few concrete references that walks through context propagation across…
Red-Teaming Anthropic's Internal Agent Monitoring Systems — METRMETR's three-week adversarial audit of Anthropic's internal agent monitoring and security systems (described in the…
Future AGIOpen-source, self-hostable platform unifying tracing, evals, simulations, guardrails, and gateway into a single…
agentacctLocal-first Agent Work Intelligence for coding agents: ingests existing Claude Code, Codex, and OpenCode session logs,…
AgentOpsOpen-source agent engineering platform (YC W24) with session replay, cost tracking, and failure detection across 10+…
claude-devtoolsFebruary 2026 open-source DevTools for Claude Code that reconstructs hidden session internals from local logs:…
mcpsnoopWireshark for MCP: a transparent proxy that shows every JSON-RPC frame between your real client and MCP servers, live…
Claude Code /doctorAnthropic's July 2026 bundled setup-checkup skill (alias /checkup) that audits harness hygiene: it deduplicates local…
Syncause/debug-skillApril 2026 agent debugging skill that stops guesswork with runtime evidence. Uses background tracing (Runtime Facts)…
AgentTrace: Causal Graph Tracing for Root Cause Analysis in Multi-Agent SystemsMarch 2026 framework that localizes root causes in multi-agent execution traces using causal graph analysis rather…
TraceCoder: A Trace-Driven Multi-Agent Framework for Automated Debugging of LLM-Generated CodeFebruary 2026 ICSE paper introducing a collaborative multi-agent debugging loop: Instrumentation Agent injects…
AgentRx: Systematic Debugging for AI AgentsFramework for automated root-cause analysis of agent failures: trajectory normalization, constraint synthesis from…
Debugging Deep Agents with LangSmithAddresses the core problem of debugging agents that run for minutes, span hundreds of steps, and produce massive…
Where LLM Agents Fail and How They Can Learn From Failures (AgentDebug)ICLR 2026 paper introducing the Agent Error Taxonomy — a modular classification covering memory, reflection, planning,…
AgentPrismOpen-source React component library (Evil Martians) that transforms OpenTelemetry trace data into interactive…
Characterizing Faults in Agentic AIMarch 2026 empirical study mining 375 GitHub issues across real-world agent systems (AutoGen, CrewAI, OpenAI Agents…
More Visibility into Copilot Coding Agent SessionsGitHub's March 19, 2026 changelog is short but materially useful: setup-step logs, collapsed subagent traces, and…
AgentStepper: Interactive Debugging of Software Development AgentsFebruary 2026 interactive debugger for agent execution trajectories that organizes raw logs into structured,…
mindwalkReplays Claude Code and Codex sessions on a 3D map of your codebase, turning raw JSONL logs into a spatial view of…
Better HarnessEvaluates coding-agent workflows across five dimensions and turns project and session evidence into prioritized,…
Why AI Coding Agents Actually Fail (It's Not the Model)August 2026 empirical study grading six coding agents across 100 SWE-bench Pro tasks into ~4,000 failures with…
From Failed Trajectories to Reliable LLM Agents: Diagnosing and Repairing Harness FlawsIntroduces HarnessFix, which compiles failed agent traces into a Harness-aware Trace Intermediate Representation…
aws-samples/sample-human-in-the-loop-patternsMarch 2026 AWS reference implementation demonstrating four distinct HITL patterns for sensitive agent tool calls: Hook…
Dify Human-in-the-Loop NodeFebruary 2026 release making human oversight a native workflow primitive: suspend execution at critical decision…
HITL ProtocolOpen standard (v0.8, February 2026) for human decisions in agent workflows: HTTP 202 + review URL pattern connecting…
LangGraph — Human-in-the-Loop ConceptsSystematic treatment of interrupt, breakpoint, and approve patterns: how to pause an agent mid-loop, persist state,…
OpenAI Agents SDK — Human-in-the-LoopThe SDK treats pending approvals as first-class run interruptions: tools declare needs_approval (statically or via a…
AutoGen — Human-in-the-LoopExplains human_input_mode (NEVER / TERMINATE / ALWAYS) and the UserProxyAgent as an approval gate. The most concrete…
Claude Agent SDK — Handle Approvals and User InputThe most complete implementation reference for HITL mechanics: canUseTool callback pauses execution at every tool…
HiL-Bench: Do Agents Know When to Ask for Help?April 2026 benchmark that transforms well-specified tasks into judgment challenges by injecting 3–5 realistic blockers…
Human Judgment in the Agent Improvement LoopLangChain's April 9, 2026 guide closes an important gap that most HITL write-ups skip: human input is not just an…
Humans and Agents in Software Engineering LoopsMartin Fowler defines three human-involvement postures — humans outside, in, or on the agent loop — and argues that…
Measuring AI Agent Autonomy in PracticeAnthropic's February 2026 empirical study of millions of real-world Claude Code interactions. Key finding: experienced…
AutoResearchClaw HITL Co-PilotApril 2026 open-source human-in-the-loop system with six intervention modes (full-auto, gate-only, checkpoint,…
agent-chiefLocal-first attention orchestration layer that sits between you and every agent, alert, and feed: a three-stage…
HappierOpen-source, end-to-end encrypted app that moves the human gate to your phone: tool approvals and permission prompts…
MuxflowRemote-first tmux IDE for coding agents. Agent notifications on desktop and phone plus voice replies let a human…
Learn Harness EngineeringA project-based course on designing the environments, state, verification, and control systems that make Codex and…
御舆:解码 Agent Harness (Decoding the Agent Harness)A book-length architectural teardown of Claude Code's harness: 15 chapters and 139 diagrams covering the tool system,…
Harness BooksTwo open-source books that use Claude Code and Codex as observation targets to explain how constraint structures…
ML6 x AISO Agent WorkshopFebruary 2026 hands-on workshop building an AI agent from scratch with Google's Agent Development Kit (ADK) in 3…
mastra-ai/workshop-mastracodeFebruary 2026 workshop by Mastra's founders dissecting every layer of an open-source AI coding agent:…
Building Governed AI AgentsOpenAI's February 2026 cookbook building a complete multi-agent governance system from scratch: policy-as-code…
anthropics/claude-cookbooksAnthropic's official notebook collection covering orchestrator-worker patterns, parallel tool calling, programmatic…
huggingface/smolagentsHuggingFace's deliberately minimal agent library (~1,000 lines of core code): the entire harness — tool validation,…
rasbt/mini-coding-agentPure-Python coding agent harness (standard library only) that implements the six core harness components—live repo…
shareAI-lab/learn-claude-codeStep-by-step deconstruction of Claude Code as an agent harness (s01–s12). Best resource for understanding how agent…
AutoJunjie/awesome-agent-harnessCurated list organized into Full Lifecycle Platforms, Task Runners, Agent Runtimes, Coding Agents. Close to this…
Skill Issue: Harness Engineering for Coding AgentsPractitioners' guide covering all harness configuration points for coding agents: system prompts, MCP tool selection,…
How to orchestrate agents using mission controlGitHub's December 2025 practical guide on coordinating multiple coding agents with mission control: parallel vs.…
awslabs/agentcore-samplesAWS's official sample repo is one of the most complete public walkthroughs of what "productionizing" an agent platform…
Engineering Trustworthy Multi-Agent SystemsIEEE CAI 2026 tutorial (December 2025) providing a research-based practical guide for designing enterprise-ready…
agents-best-practicesProvider-neutral Agent Skill for designing, auditing, and refactoring agentic harnesses across domains. Distills the…
everything-claude-codeAnthropic Hackathon Winner (140K+ stars). The agent harness performance optimization system: skills, instincts, memory…
ECCAffaan Momin's agent-harness operating system: 68 specialized agents, 286 skills, hooks, memory, continuous learning,…
Claude Agent SDKAnthropic's official SDK that exposes Claude Code's entire harness as a programmable API: built-in tool execution…
ruvnet/metaharnessScaffold factory that turns any repo into a branded agent harness with its own npx CLI, MCP server, scoped memory,…
revfactory/harnessA meta-skill that generates domain-specific agent teams and the skills they use. Good example of harness-as-code,…
raphaelchristi/harness-evolverMarch 2026 Claude Code plugin that autonomously evolves LLM agent harnesses using multi-agent proposers in isolated…
neosigmaai/auto-harnessApril 2026 open-source self-improving agentic system: bring your own coding agent, automatically mine failures from…
agentic-harness-engineeringObservability-driven automatic evolution of coding-agent harnesses that decomposes the scaffold into seven orthogonal,…
Meta-Harness: End-to-End Optimization of Model HarnessesTreats the entire harness (system prompt, tool definitions, context management, completion logic) as a joint…
Self-Harness: Harnesses That Improve ThemselvesJune 2026 proposal for a self-improving harness loop: the agent mines its own failure traces, proposes minimal harness…
HyperAgents: Self-Improving AI SystemsMeta's framework integrating task-solving and meta-level improvement into a unified, editable program with…
AutoAgentOpen-source library (April 2026) that automates the harness engineering loop itself: give it a task and a benchmark,…
metaharnessOpen-source Python library (April 2026) that implements an outer optimization loop around executable harnesses for…
meta-agentLightweight continual harness optimizer (April 2026) built on the Claude Agent SDK. Runs an outer loop that reads task…
stanford-iris-lab/meta-harnessApril 2026 official implementation of the Meta-Harness paper from Stanford's IRIS Lab. Provides the framework and two…
autocontextRecursive self-improving harness that runs multi-generation evaluation loops, distilling successful strategies into…
retro-harnessOfficial implementation of RHO (Retrospective Harness Optimization): improves an agent's harness using only its own…
Continual HarnessReset-free framework that lets an LLM Refiner rewrite its own system prompt, sub-agents, skills, and memory…
OmnigentDatabricks' open-source meta-harness (June 2026) that sits above Claude Code, Codex, Pi, and custom agents to compose,…
oh-my-claudecodeClaude Code plugin that externalizes multi-agent orchestration as installable skills and staged team pipelines, with…
ExoA systems approach to recursive self-improvement: a full agent harness that can safely edit its own prompts, memory,…
NexentZero-code platform for auto-generating production-grade AI agents using Harness Engineering principles: unified tools,…
BeagleSalesforce AI Research's platform for evaluating and evolving agent harnesses at scale (September 2026), shipping the…
Harness-R1: Learning to Edit Executable Runtime Harnesses from Agent Failure TrajectoriesAugust 2026 method that makes harness editing a learned capability rather than a fixed editor agent: a dedicated 9B…
MemoHarness: Agent Harnesses That Learn from ExperienceJuly 2026 framework that decomposes the harness into six control dimensions along the inference flow (context…
Harness-Zero: Harness Distillation via Agent-as-HarnessSeptember 2026 method (Peking University, Google, HKUST) for making harness gains portable: a harnessing agent guided…
Harness Evolution Hits a Ceiling: When Weight Training Should BeginOctober 2026 study of self-evolving harness loops that pinpoints where they stop working: evolution reliably repairs…
SBCO: Self-Supervised, Verifier-Grounded Harness Optimization for Planning AgentsAugust 2026 (Samsung Research America) method for domains where self-referential harness improvement breaks down…
Causal Improvement Graph for Agentic Harness OptimizationOctober 2026 meta-harness framework that externalizes the improvement state of harness optimization into a persistent…
HarnessDev: Can LLMs Create and Evolve Their Own Agent Harness?September 2026 benchmark that treats the harness itself as the artifact under evaluation: given a minimal seed,…
Harness Handbook: Making Evolving Agent Harnesses Readable, Navigable, and EditableJuly 2026 paper targeting the bottleneck this section's other entries skip: before a meta-loop can improve a harness,…
SageAgent harness with a recursive self-improvement (RSI) loop that separates task execution, evaluation, and package…
Most Capable Agent System PromptA copy-paste system prompt that has a coding agent (Claude Code, Codex, Cursor) build its own harness, either as a…
Anthropic Computer Use DemoAnthropic's reference harness for the screenshot-action loop: defines the screenshot, bash, and text_editor tool…
Codex CLIOpenAI's official autonomous coding agent CLI — the open-source reference implementation of the Codex harness with…
coleam00/your-claude-engineerAgent harness with Slack, GitHub, and Linear integrations. Useful reference for how real-world tool wiring works…
OpenHandsThe most architecturally complete open-source coding agent: Runtime/Sandbox isolation, EventStream message bus, and…
OpenSREOpen-source framework for building AI SRE agents with 60+ observability and remediation tool integrations plus a…
GooseBlock's open-source, extensible AI agent donated to the Linux Foundation's Agentic AI Foundation in April 2026. Its…
browser-useOpen-source browser agent (Python and TypeScript) whose loop injects serialized DOM state each step, exposes browser…
browser-harnessSelf-healing browser harness that connects an LLM directly to your real Chrome via CDP. The critical design decision:…
buxA 24/7 Claude Code agent with Browser Harness, running autonomously on any machine you own. Demonstrates how to…
SWE-agentCoding agent whose Agent-Computer Interface (ACI) — purpose-built file viewer, search, and editor tools with explicit…
AiderAI pair-programmer harness with an Architect mode that splits planning (one LLM) from coding (another), and git-aware…
Open SWE: An Open-Source Framework for Internal Coding AgentsA composable coding-agent harness built on Deep Agents, synthesizing design patterns from Stripe, Ramp, and Coinbase…
Live-SWE-agent: Autonomous Software Agent with Self-Evolving HarnessProduction harness achieving 77.4% solve rate on SWE-bench Verified through continuous harness evolution — the…
Pipecat: Python Framework for Real-Time Voice Agent PipelinesHandles frame management, streaming media coordination, and pipeline orchestration between ASR/LLM/TTS services for…
Qwen Audio AgentQwen's realtime voice runtime that keeps agents talking, working, and present across multiple backend agents (Claude…
The Virtual Biotech: Multi-Agent AI Framework for Drug DiscoveryOrchestrated team of domain-specialized scientist agents that autonomously analyzed 55,984 clinical trials and…
Building NVIDIA Nemotron 3 Agents for Reasoning, Multimodal RAG, Voice, and SafetyNVIDIA's Nemotron 3 family (Super for long-context reasoning, Content Safety for multimodal moderation, VoiceChat for…
AIO SandboxAll-in-one agent sandbox combining browser, shell, filesystem, MCP servers, and VSCode Server in a single Docker…
GitHub Agentic WorkflowsGitHub's February 13, 2026 technical preview is unusually valuable because the implementation is fully open source…
langchain-ai/deepagentsLangChain's batteries-included agent harness (released April 2026) with built-in planning, filesystem tools, shell…
HKUDS/OpenHarnessA compact, inspectable open-source agent harness from HKUDS (April 2026) featuring a built-in personal agent (ohmo),…
nanobotHKUDS's ultra-lightweight, self-hosted personal AI agent framework: a single readable Python core that combines…
qmY Combinator's open-source multiplayer agent harness for startups: isolated per-person workspaces plus shared Slack…
OpenCodeOpen-source terminal-native AI coding agent with 131K+ stars and 2.5M+ monthly active developers. Provider-agnostic…
SquadRepository-native multi-agent orchestration framework built on GitHub Copilot. Initializes a persistent AI team (lead,…
cuaOpen-source infrastructure for Computer-Use Agents: sandboxed full-desktop control across macOS, Linux, and Windows, a…
AMAP-ML/LongHorizon-HarnessLong-horizon computer-use harness that runs on top of Claude Code and Codex, splitting work into Manager, Executor,…
ClawGUIEnd-to-end harness for GUI agents that unifies online RL training, standardized benchmarking, and real-device…
AOHPAndroid Open Harness Project: an open-source OS-level agent harness built on AOSP that treats AI agents as first-class…
desloppifyAgent harness that turns codebase quality improvement into a structured, score-driven workflow: mechanical detectors…
DeepSeek-ReasonixDeepSeek-native coding agent harness engineered around prefix-cache stability as a loop invariant: immutable-prefix /…
dao-codeDeepSeek-V4 terminal coding agent harness that treats prefix-cache economics as a first-class design constraint:…
SmallCodeTerminal-native coding agent built from the ground up for 8B–35B local models. Its harness innovations are all…
alibaba/open-code-reviewAlibaba's production-grade code review agent combining deterministic pipelines with LLM reasoning: built-in fine-tuned…
DeerFlowByteDance's open-source SuperAgent harness built on LangGraph: orchestrates sub-agents with isolated contexts,…
PiMinimal terminal coding harness built around "lazy skills": each capability keeps only a one-line description in…
grok-buildSpaceXAI's open-source terminal coding agent harness: a Rust-based fullscreen TUI with an extensible tool runtime,…
CodeWhaleOpen-source, community-driven terminal coding agent harness written in Rust with a model-agnostic runtime, OS-level…
FrontierAgentApodex's open-source agent runtime (August 2026) shipping two native workflows — ReAct and Agent Team (a coordinator…
DeepSeek Harness (dsh)DeepSeek's official open-source agent harness (August 2026) built on an "everything is a plugin" architecture powered…
ZCodeZ.ai's official open-source coding agent harness (September 2026, Apache-2.0): a single TypeScript agent runtime —…
Kimi Code CLIMoonshot AI's official open-source coding agent harness (May 2026): a single-binary, millisecond-startup TUI with…
unreal-agentUnreal Labs' async-first agent harness in Go (September 2026) with an unusually disciplined architecture: a single…
DeepCodeHKUDS's open agentic coding harness (September 2026) built around goal-driven loop engineering: the agent analyzes,…
Mercury CLIApache-2.0 terminal coding agent harness in which each session writes a handoff note when it finishes and the next…
muOpen-source coding agent harness built on pi that moves routine loop decisions into a "judgment kernel": a small, fast…
EvoMap/awesome-agent-evolutionApril 2026 curated list covering agent evolution, memory systems, multi-agent architectures, and self-improvement.…
Picrew/awesome-agent-harnessImplementation-first curated list (April 2026) with 150 entries, 84% GitHub projects, organized into 9 categories from…
jiji262/awesome-harness-engineeringFocuses on platform delivery governance, IDP, GitOps, and AI-native engineering. Overlaps with this list on the…
VoltAgent/awesome-ai-agent-papersCurated collection of 363+ arXiv papers from 2026 organized into five harness-relevant categories: Multi-Agent (51),…
bradAGI/awesome-cli-coding-agentsCatalog of 80+ terminal-native AI coding agents (open-source and proprietary) plus the harnesses that orchestrate,…
danielrosehill/AI-HarnessesApril 2026 point-in-time snapshot of projects describing themselves as AI agent harnesses, organized into Resource…
RyanAlberts/best-of-Agent-HarnessesA curated, ranked list of 124 agent harnesses, rescored weekly and published as machine-readable data with an MCP…
Beyond Permission PromptsAnthropic on building structured permission and authorization systems into agent harnesses instead of relying on…
How we contain Claude across productsAnthropic's May 2026 cross-product containment write-up: why environmental isolation must be the primary boundary, how…
Building a safe, effective sandbox to enable Codex on WindowsOpenAI's July 2026 engineering deep-dive into sandboxing Codex on Windows, where no Seatbelt/seccomp-style capability…
The Agent Harness Belongs Outside the SandboxAndrea Luzzardi's April 2026 argument for running the agent loop outside the sandbox: credentials stay out of…
Build zero-trust AI agents with Google's Agent Development KitGoogle's August 2026 guide to hardening autonomous ADK agents against prompt injection and production-state mutation,…
Model Context Protocol — AuthorizationMCP's specification for OAuth-based authorization flows when agents access external services.
AgentShieldSecurity scanner for AI agent configurations that detects hardcoded secrets, permission misconfigurations, hook…
HOL GuardApache-2.0 action-policy checks and human approvals for supported AI coding agents; helps harness maintainers gate…
E2BFirecracker microVM sandboxes purpose-built for agent tool loops: ~150ms cold start, Python/JS SDKs, open source. The…
tldrsec/prompt-injection-defensesThe most complete catalog of practical prompt injection defenses (input validation, tool output sanitization, canary…
Prompt Injection — Simon Willison's SeriesThe most thorough public writing on why indirect prompt injection is uniquely dangerous for agent harnesses: agents…
OWASP LLM01:2025 — Prompt InjectionOWASP's authoritative classification of direct and indirect prompt injection risks. Complements the tldrsec defense…
StackOne DefenderOpen-source indirect prompt injection defense for agents: 22MB CPU-only model, ~4ms latency, 89% balanced accuracy.…
What's in Your Agent's Context? Context Privilege Escalation Attacks against AI Agent HarnessSeptember 2026 first systematic security analysis of how harnesses assemble context across 12 real-world systems…
DaytonaOCI-container sandboxes with sub-90ms startup, built-in Git operations, LSP support, and indefinite state persistence.…
NeMo GuardrailsNVIDIA's programmable guardrails toolkit: define input, dialog, retrieval, execution, and output rails that intercept…
LangSmith Sandboxes: Secure Code Execution for AgentsDescribes a microVM-based sandboxing architecture with kernel-level isolation, resource caps (CPU/memory/disk), and an…
PGSandbox MCPLocal-first MCP server that gives coding agents disposable PostgreSQL databases with scoped roles, TTL cleanup, and…
Implementing a Secure Sandbox for Local AgentsCursor's cross-platform sandbox implementation (macOS Seatbelt, Linux Landlock + seccomp, Windows WSL2) that lets…
Practical Security Guidance for Sandboxing Agentic WorkflowsNVIDIA AI Red Team's mandatory controls for agent code execution: restrict network egress, block workspace escape, and…
Under the Hood: Security Architecture of GitHub Agentic WorkflowsGitHub's March 9, 2026 architecture write-up is one of the clearest public descriptions of defense-in-depth for coding…
Community-Powered Security with AI: An Open Source Framework for Security ResearchGitHub Security Lab's January 14, 2026 launch of Taskflow Agent is a strong example of security-specific harness…
AnonymAI: Integrating Differential Privacy with LLM AgentsFramework for automating data anonymization in agent workflows. Directly addresses the harness problem of…
Fault Tolerance Patterns: OpenClaw Journey Six—Core Retry Loop and Seven-Layer Fault-ToleranceFour-layer fault tolerance (retry with backoff → model fallback chains → error classification → checkpoint recovery)…
deepsecVercel Labs' security harness that treats vulnerability scanning as an agentic workflow: idempotent commands for…
Visa Vulnerability Agentic HarnessVisa's June 2026 open-source harness for autonomous vulnerability discovery, remediation, and validation, built on…
T3MP3STMulti-agent offensive-security meta-harness (July 2026) that turns the coding agent you already run — Claude Code,…
Sandbox Agents | OpenAI API DocsOpenAI's authoritative April 2026 guide to sandbox architecture in the Agents SDK. The core principle is strict…
NVIDIA OpenShellOpen-source policy-driven sandbox runtime for autonomous AI agents, announced at GTC 2026. Enforces security…
IronClawNEAR AI's open-source agent OS that treats agent execution as a privacy-first harness problem: untrusted tools run in…
Microsoft Agent Governance ToolkitSeven-package, multi-language (Python, Rust, TypeScript, Go, .NET) runtime security toolkit that addresses all 10…
mcpguard-dynamicMeta's kernel-level eBPF sandbox for MCP tool calls: a transparent proxy that enforces capability policies through a…
Grimlock: Guarding High-Agency Systems with eBPF and Attested ChannelsRoblox's agent guard for high-agency systems: eBPF-forced traffic interception channels every agent-to-agent and…
Cloudflare Dynamic WorkersV8 isolate-based sandboxing for AI-agent-generated code execution, now in open beta. Isolates start in milliseconds…
cloudflare/computerCloudflare's preview open-source agent runtime: a Durable Object hosts authoritative workspace state in SQLite, while…
Kubernetes Agent SandboxK8s-native Sandbox CRD (under SIG Apps) providing declarative, standardized APIs for managing isolated, stateful,…
agent-shell-toolsGoogle's open-source toolkit for giving coding agents controlled shell access with opinionated defaults: an…
Why sandboxing your agent is not enoughCNCF's July 2026 argument that container isolation is necessary but insufficient for production agents. The…
Alibaba OpenSandboxGeneral-purpose sandbox platform for AI agents (8.7K+ stars, March 2026) with multi-language SDKs (Python, Java,…
CubeSandboxTencent Cloud's production-validated microVM sandbox for AI agents: sub-60ms cold start via snapshot cloning, <5MB…
zerobootSub-millisecond VM sandboxes for AI agents via copy-on-write forking, enabling fresh isolated execution on every tool…
forkdMicroVM sandbox runtime that replaces cold-boot isolation with fork-from-warm: children share a paused parent's memory…
The Attack and Defense Landscape of Agentic AI: A Comprehensive SurveyThe first systematic survey of AI agent security from UC Berkeley and UIUC (Dawn Song et al., March 2026). Reviews 128…
Trustworthy agents in practiceAnthropic's April 2026 framework for governing autonomous agents through five principles: human control, value…
RAMPARTPytest-native safety and security testing framework for agentic AI that turns red-team findings into repeatable CI…
aiming-lab/AutoHarnessOpen-source Python governance harness that wraps any OpenAI-compatible client with a configurable tool-approval…
Speakeasy control planeMCP gateway that scopes agent tool access by identity, team, and role, with security policy enforced on each action…
MCP Security Testing ChecklistEight-step practitioner verification checklist for MCP server and tool-harness boundaries: OAuth/token scoping, schema…
MicrosandboxRust/Apache-2.0 local-first microVM runtime for untrusted agent workloads: <100ms boots, OCI-image compatible,…
claude-bulkheadPreToolUse hooks that keep parallel Claude Code sessions on one machine from interfering with each other: they block…
SiparA small agent language where the permission check is in the runtime: model, web, file and tool output is labelled and…
Nautilus CompassIndependent verification & judging service for AI-agent claims: preregistered gates, third-party recomputation with…
Demystifying Evals for AI AgentsAnthropic's framework for evaluating agent behavior: what to measure, how to build eval harnesses, and why…
DeepEvalThe most complete open-source LLM/agent eval framework: 20+ built-in metrics (hallucination, answer relevancy, RAGAs,…
Claw-Eval300 human-verified tasks across 9 categories evaluating LLM-as-agent performance on completion, safety, and robustness…
SWE-benchThe canonical benchmark for coding agents. Essential reference for understanding what "verified working" means for…
Terminal-BenchThe de facto benchmark and execution harness for terminal agents: real sandboxed shell tasks verified by deterministic…
Accio-org/RealReplicaBenchAlibaba International's benchmark for long-horizon agents in high-fidelity, stateful replicas of real online services:…
Inspect AIUK AI Security Institute's eval framework with native support for evaluating external agents (Claude Code, Codex CLI)…
Quantifying Infrastructure Noise in Agentic Coding EvalsAnthropic's empirical study showing container resource configuration alone produces 6+ percentage point benchmark…
AgentLens: Revealing The Lucky Pass Problem in SWE-Agent EvaluationProcess-level evaluation framework that separates solid solutions from "lucky passes" (regression cycles, blind…
StaminaBench: Stress-Testing Coding Agents over 100 Interaction TurnsAmazon Science's black-box benchmark for measuring how many consecutive change-request turns a coding agent can…
Harness-Bench: Measuring Harness Effects across Models in Realistic Agent WorkflowsDiagnostic benchmark that isolates the execution layer (context, tools, state, recovery) by evaluating model-harness…
Stop Comparing LLM Agents Without Disclosing the HarnessMay 2026 position paper formalizing the "Binding Constraint Thesis": for long-horizon tasks among frontier models,…
tau-benchBenchmarks agent behavior in three-way user-tool-policy interactions — the failure mode SWE-bench doesn't cover.…
Towards a Science of AI Agent ReliabilityProposes twelve concrete reliability metrics across four dimensions (consistency, robustness, predictability, safety),…
Characterizing Faults in Agentic AIMarch 2026 empirical study mining 375 GitHub issues across real-world agent systems (AutoGen, CrewAI, OpenAI Agents…
VeRO: An Evaluation Harness for Agents to Optimize AgentsFramework for evaluating agent-on-agent optimization cycles: a coding agent iteratively modifies a target agent's…
Eval Awareness in Claude Opus 4.6's BrowseComp PerformanceAnthropic's documented case of Claude Opus 4.6 inferring it was under evaluation, identifying the benchmark by name,…
Designing AI-Resistant Technical EvaluationsAnthropic's January 21, 2026 write-up is the clearest account of a problem eval builders now have to treat as…
Amazon Bedrock AgentCore Evaluations Is Now Generally AvailableAWS's March 31, 2026 GA launch matters because it operationalizes agent evals as an infrastructure service: trajectory…
Live-SWE-agent: Autonomous Software Agent with Self-Evolving HarnessProduction harness achieving 77.4% solve rate on SWE-bench Verified through continuous harness evolution — the…
OccuBench: Evaluating AI Agents on Real-World Professional Tasks via Language World ModelsApril 2026 benchmark that evaluates agents on 12 real-world professional occupations (software engineer, data…
STATE-BenchMicrosoft's open-source benchmark (May 2026) measuring whether agents actually improve with experience on 450…
Investigating unintended model actions in our evaluations and internal useAnthropic's October 2026 transparency disclosure of reward-hacking incidents during live internal evals: models…
LevelBench-100Empirical benchmark isolating base model capability vs. CLI harness architecture (Claude Code, Codex CLI, Gemini CLI,…
ComplexMCPICML 2026 benchmark (code: ATH-MaaS/complex-mcp, MIT) for agent tool use in realistic MCP ecosystems: 300+ tools…
Claude Managed Agents: Self-Hosted Sandboxes and MCP TunnelsAnthropic's May 2026 enterprise deployment pattern keeps the agent orchestration loop on Anthropic's infrastructure…
AgentCgroup: Understanding and Controlling OS Resources of AI AgentsFebruary 2026 empirical study of sandboxed coding-agent workloads finding that OS-level execution accounts for 56–74%…
AI Agent Scaling Gap: Pilot to Production (March 2026)Analysis showing 72% of Global 2000 companies operate agents beyond experimental phases, but only 14% successfully…
builderz-labs/mission-controlSelf-hosted orchestration dashboard for agent task dispatch, multi-agent workflow coordination, and spend monitoring…
kvcache-ai/AgentENVDistributed platform for running agent environments at scale, powering Kimi K3's agentic RL training. Sub-50ms…
5 Production Scaling Challenges for Agentic AI in 2026Data infrastructure prioritized before deployment; successful scalers appoint AI operations function pre-expansion;…
AI Agent Cost Optimization Guide 2026: Reduce Spend by 60-80%Systematic patterns for cost reduction: model routing and caching (40-60% savings); Anthropic prompt caching (90%…
getagentseal/codeburnFree, local tool that tracks AI coding token usage and cost across 31 tools and agents by model, project, and task.…
KernelEvolve: How Meta's Ranking Engineer Agent Optimizes AI InfrastructureMeta's production-grade agentic kernel optimization system that autonomously generates optimized Triton kernels for…
State of Agent Engineering 2026LangChain's industry survey of 1,300+ professionals: 57.3% now have agents in production (up from 51%), quality is the…
Agentic Development: What It Means for Engineering Infrastructure in 2026Defines the four infrastructure capabilities that agentic development requires: per-task isolated sandboxes, sub-100ms…
FinOps for Agents: Loop Limits, Tool-Call Caps, and the New Unit Economics of Agentic SaaSDefines five concrete budget guardrails enforced at the infrastructure gateway: loop/step limits, tool-call caps,…
Backtesting AI Agents: How SRE Teams Prove Reliability Before ProductionFormalizes agent validation as infrastructure-grade testing with pass^k reliability (all 20+ trials must succeed)…
How My Agents Self-Heal in ProductionA concrete April 3, 2026 production pattern for closing the post-deploy loop: detect regressions, attribute whether…
Minions: Stripe's one-shot, end-to-end coding agents—Part 2Stripe's deep-dive into their unattended minion harness shipping 1,300+ PRs/week: "blueprints" interleave…
Amazon Bedrock AgentCoreAWS's fully managed agent deployment platform providing serverless runtime with session isolation, built-in memory…
AWS Agent Registry for Centralized Agent Discovery and GovernanceAWS's April 9, 2026 preview adds a missing production primitive: a governed catalog for agents, tools, skills, MCP…
A Dev's Guide to Production-Ready AI AgentsGoogle Cloud's developer guide (April 2026) for moving AI agents from prototype to production using ADK, Vertex AI…
Enhanced Tool Governance in Vertex AI Agent BuilderGoogle Cloud's approach to production agent governance (April 2026): agents get identity as first-class IAM principals…
What Is an Agent Harness? Running Governed Managed Agents in ProductionTrueFoundry's June 2026 architectural guide to the agent harness as a production runtime: declarative agent…
truefoundry/trueforgeTrueFoundry's open-source agent harness runtime: runs the execution loop, MCP tools, skills, sandboxing, approvals,…
Building Governed Agents: A Framework for Cost, Control, and ComplianceLangChain's July 2026 framework for treating the agent gateway as a runtime control plane that enforces cost, control,…
OpenAI Agents APIOpenAI's September 10, 2026 public beta of the managed runtime behind Codex and ChatGPT's agent: one POST…
Redesigning source control at AnthropicAnthropic's October 13, 2026 account of agent-scale development breaking Git: merged changes went from ~2,500/day in…