Skip to content

Awesome list for AI agent harness engineering: tools, patterns, evals, memory, MCP, permissions, observability, and orchestration.

Round 11: add 3 latest harness engineering resources

ai-boostea06148e471f
diff --git a/README.md b/README.md
index e9e9a55..275b424 100644
--- a/README.md
+++ b/README.md
@@ -116,6 +116,7 @@ Harness components organized by the problem they solve, not by vendor.
- [Choosing the Right Multi-Agent Architecture](https://blog.langchain.com/choosing-the-right-multi-agent-architecture/) — Decision framework for four multi-agent patterns (subagents, skills, handoffs, router) with concrete performance data: subagents process 67% fewer tokens than skills in multi-domain scenarios because context isolation prevents cross-domain bloat. The five-dimension matching table (distributed development, parallelization, multi-hop, user interaction, latency) is the most actionable published guide for deciding when a topology change — not a model change — is the right lever for a performance problem.
- [Multi-Agent Workflows Often Fail. Here's How to Engineer Ones That Don't.](https://github.blog/ai-and-ml/generative-ai/multi-agent-workflows-often-fail-heres-how-to-engineer-ones-that-dont/) — GitHub's February 24, 2026 distillation of a failure pattern most harnesses eventually rediscover: multi-agent systems behave like distributed systems, so every handoff needs typed schemas, constrained action schemas, and explicit boundary validation. Worth including because it turns "add more agents" from a vibe into an interface design problem you can actually reason about.
- [Effective Harnesses for Long-Running Agents](https://www.anthropic.com/engineering/effective-harnesses-for-long-running-agents) — Anthropic's pattern for maintaining agent progress across multiple context windows: an initializer agent sets up the environment once and hands off to a coding agent that makes incremental progress each session. The structured handoff mechanism — feature lists, git commits, and test gates as cross-session state — is the reference design for any harness where a task exceeds a single context window and naïve restarts lose accumulated progress.
+- [Task-Adaptive Multi-Agent Orchestration (AdaptOrch)](https://arxiv.org/abs/2602.16873) — February 2026 framework that dynamically selects orchestration topology (parallel, sequential, hierarchical, or hybrid) based on task dependency graphs rather than fixed pipeline architecture. Demonstrates that topology choice is a harness-level lever that can improve performance 12–23% over model selection alone.
### Context Delivery & Compaction
@@ -143,6 +144,7 @@ Harness components organized by the problem they solve, not by vendor.
- [AutoHarness: Improving LLM Agents by Automatically Synthesizing a Code Harness](https://arxiv.org/abs/2603.03329) — Google DeepMind technique that uses code synthesis to auto-generate runtime constraint harnesses from tool schemas and task specifications. Gemini-2.5-Flash + AutoHarness outperforms Gemini-2.5-Pro and GPT-5.2-High on TextArena games by eliminating illegal moves through learned harness policies. Shifts constraint enforcement from static (schema validation) to dynamic (synthesized code guards) — a reference pattern for learning-based behavioral guardrails.
- [Scaling Parallel Tool Calling for Efficient Deep Research](https://arxiv.org/abs/2602.07359) — February 2026 analysis of how parallel tool calling reduces latency in multi-step agent workflows. Demonstrates that concurrent tool execution (rather than sequential observe→act loops) is the key efficiency lever for deep-research harnesses where each step may invoke search, browse, and compute tools simultaneously. Essential for designing low-latency agent loops without sacrificing reasoning depth.
- [EigentSearch-Q+](https://arxiv.org/abs/2604.07927) — April 2026 framework for deep-research agents using dedicated reasoning tools (plan_next_searches, select_query_and_search, extract_relevant_details, analyze_search_progress) that externalize intermediate decisions as typed tool arguments. Inspired by Anthropic's think-tool paradigm, Q+ makes cognitive scaffolding explicit and auditable — bridging classic information-retrieval strategies with structured model-driven tool invocations.
+- [Design Patterns for Deploying AI Agents with Model Context Protocol](https://arxiv.org/abs/2603.13417) — March 2026 field report from an enterprise MCP deployment identifying three protocol-level gaps that break production: missing identity propagation (who is the request for?), absent adaptive tool budgeting, and unstructured error semantics. The concrete mitigation patterns — JWT-enriched tool calls, per-tool timeout contracts, and standardized error-action mappings — are essential before betting on MCP as your primary tool-integration layer.
### Skills & MCP
@@ -272,6 +274,7 @@ Harness components organized by the problem they solve, not by vendor.
- [Human Judgment in the Agent Improvement Loop](https://blog.langchain.com/human-judgment-in-the-agent-improvement-loop/) — LangChain's April 9, 2026 guide closes an important gap that most HITL write-ups skip: human input is not just an approval gate at execution time, it's also supervision for improving prompts, tools, memory, and evaluators over time. Useful because it treats expert review as a structured data source for harness evolution rather than a one-off manual checkpoint.
- [Humans and Agents in Software Engineering Loops](https://martinfowler.com/articles/exploring-gen-ai/humans-and-agents.html) — Martin Fowler defines three human-involvement postures — humans outside, in, or on the agent loop — and argues that "humans on the loop" (maintaining the harness rather than reviewing individual outputs) is the only approach that scales with agent throughput. The "agentic flywheel" section — where agents are directed to evaluate results and recommend harness improvements — is the clearest articulation of how HITL evolves from a gate into a feedback mechanism.
- [Measuring AI Agent Autonomy in Practice](https://www.anthropic.com/news/measuring-agent-autonomy) — Anthropic's February 2026 empirical study of millions of real-world Claude Code interactions. Key finding: experienced users shift from per-action approval (20% auto-approve when new) to intervention-only oversight (40% auto-approve at 750+ sessions), and agent-initiated clarification stops grow faster than human interruptions as task complexity increases. The most data-grounded reference for designing adaptive permission models that scale with user trust.
+- [AutoResearchClaw HITL Co-Pilot](https://github.com/aiming-lab/AutoResearchClaw) — April 2026 open-source human-in-the-loop system with six intervention modes (full-auto, gate-only, checkpoint, step-by-step, co-pilot, custom), SmartPause confidence-driven dynamic suspension, and Intervention Learning from human corrections. The cost-guardrail system — aborting runs that exceed budget thresholds — makes it a practical reference for production HITL where human time is as constrained as agent compute.
---