Skip to content

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

Add TomasuLLM (arXiv 2609.38201) to Agent Loop

ai-boost13a871541d0f
diff --git a/README.md b/README.md
index 8aa3051..a9178fe 100644
--- a/README.md
+++ b/README.md
@@ -126,6 +126,7 @@ Harness components organized by the problem they solve, not by vendor.
- [statewright](https://github.com/statewright/statewright) — State machine guardrails that constrain which tools an agent can call in each phase of a workflow, turning open-ended loops into deterministic state transitions. The research result is striking: local models went from 2/10 to 10/10 passing on a SWE-bench subset purely by shrinking the tool space, proving that loop structure — not model size — is the binding constraint. ![Stars](https://img.shields.io/github/stars/statewright/statewright?style=flat-square&label=★&color=yellow)
- [Introducing dynamic workflows in Claude Code](https://claude.com/blog/introducing-dynamic-workflows-in-claude-code) — Anthropic's May 2026 introduction to dynamic parallel subagent orchestration: Claude generates JavaScript orchestration scripts that fan out work to tens or hundreds of parallel subagents with adversarial verification, converging on answers for tasks like the 750k-line Bun Zig-to-Rust port. The key harness insight is that the plan lives in executable code rather than the model's context window, scaling the agent loop to work that would otherwise exceed a single context window.
- [AgentSPEX](https://github.com/ScaleML/AgentSPEX) — UIUC's open-source specification and execution language for LLM-agent workflows: declarative YAML with typed steps, branching, loops, and explicit state management, backed by a Docker sandbox with 50+ MCP tools, checkpointing, and trajectory logging. A concrete reference for turning ad-hoc agent loops into version-controlled, reproducible harness artifacts. ![Stars](https://img.shields.io/github/stars/ScaleML/AgentSPEX?style=flat-square&label=★&color=yellow)
+- [TomasuLLM: Out-of-Order Speculative Execution for LLM Agents](https://arxiv.org/abs/2609.38201) — September 2026 paper that ports speculative execution from processor design to the agent loop: the harness predicts future tool calls, runs them early in isolated copy-on-write sandboxes, and commits results in the original trajectory order only after validation — attacking the "observation stall" where long compiles and test suites leave the agent idle. 1.31× speedup on SWE-bench Verified and 1.35× on Terminal-Bench 2.0 with zero false accepts across 4,010 audited commit records make it the first credible demonstration that the sequential observe→act loop leaves real latency on the table, and that loop-level parallelism is safe when validation and commit ordering are harness-enforced rather than model-decided.
### Planning & Task Decomposition