# MOSS: Self-Evolution through Source-Level Rewriting in Autonomous Agent Systems arXiv: https://arxiv.org/abs/2605.22794 ยท HTML: https://mlsanity.com/paper/2605.22794/ ## Key facts - Cited by 13 papers in the map (data through 14 Oct 2026). - Placed in General LLM agent frameworks and benchmarks, one of 16 directions in LLM agents and tool use. - Builds on 23 of its 31 references that are in the map. - First posted on arXiv 21 May 2026. ## Placed in - LLM agents and tool use / [General LLM agent frameworks and benchmarks](https://mlsanity.com/agents/general-llm-agent-frameworks-and-benchmarks/) ## Builds on (most cited first, 23 of 23) - [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023) - [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022) - [Toolformer: Language Models Can Teach Themselves to Use Tools](https://mlsanity.com/paper/2302.04761/) (2023) - [A Survey on Large Language Model based Autonomous Agents](https://mlsanity.com/paper/2308.11432/) (2023) - [AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation](https://mlsanity.com/paper/2308.08155/) (2023) - [ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIs](https://mlsanity.com/paper/2307.16789/) (2023) - [MetaGPT: Meta Programming for A Multi-Agent Collaborative Framework](https://mlsanity.com/paper/2308.00352/) (2023) - [CAMEL: Communicative Agents for "Mind" Exploration of Large Language Model Society](https://mlsanity.com/paper/2303.17760/) (2023) - [ChatDev: Communicative Agents for Software Development](https://mlsanity.com/paper/2307.07924/) (2023) - [DSPy: Compiling Declarative Language Model Calls into Self-Improving Pipelines](https://mlsanity.com/paper/2310.03714/) (2023) - [DeepSeek-V3.2: Pushing the Frontier of Open Large Language Models](https://mlsanity.com/paper/2512.02556/) (2025) - [Large Language Model based Multi-Agents: A Survey of Progress and Challenges](https://mlsanity.com/paper/2402.01680/) (2024) - [GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning](https://mlsanity.com/paper/2507.19457/) (2025) - [Automated Design of Agentic Systems](https://mlsanity.com/paper/2408.08435/) (2024) - [Darwin Godel Machine: Open-Ended Evolution of Self-Improving Agents](https://mlsanity.com/paper/2505.22954/) (2025) - [Meta-Harness: End-to-End Optimization of Model Harnesses](https://mlsanity.com/paper/2603.28052/) (2026) - [SkillClaw: Let Skills Evolve Collectively with Agentic Evolver](https://mlsanity.com/paper/2604.08377/) (2026) - [A Self-Improving Coding Agent](https://mlsanity.com/paper/2504.15228/) (2025) - [Claw-Eval: Towards Trustworthy Evaluation of Autonomous Agents](https://mlsanity.com/paper/2604.06132/) (2026) - [Hyperagents](https://mlsanity.com/paper/2603.19461/) (2026) - [EvoAgentX: An Automated Framework for Evolving Agentic Workflows](https://mlsanity.com/paper/2507.03616/) (2025) - [GenericAgent: A Token-Efficient Self-Evolving LLM Agent via Contextual Information Density Maximization (V1.0)](https://mlsanity.com/paper/2604.17091/) (2026) - [From Procedural Skills to Strategy Genes: Towards Experience-Driven Test-Time Evolution](https://mlsanity.com/paper/2604.15097/) (2026) ## Built on by (most cited first, 13 of 13) - [TTHE: Test-Time Harness Evolution](https://mlsanity.com/paper/2607.08124/) (2026) - [AutoDesign: Meta-Harness Optimization for Long-Horizon Agentic Design](https://mlsanity.com/paper/2608.13560/) (2026) - [EvolveNet: Collaborative Harness Evolution for Agent Self-Improvement](https://mlsanity.com/paper/2608.04968/) (2026) - [Code Is the Body: Agent-Owned Software Bodies for Recursive Evolution and Descent](https://mlsanity.com/paper/2607.28691/) (2026) - [Runtime-Independent Persistent Agents: Preserving Identity, Memory, and Code Across Models, Harnesses, and Servers](https://mlsanity.com/paper/2609.00546/) (2026) - [MetaRSI / RSI2: A Meta-Recursive Self-Improving System for Recursive Self-Improving Systems Themselves](https://mlsanity.com/paper/2609.06396/) (2026) - [A Theory of Reliable Self-Evolution for Agent Harnesses](https://mlsanity.com/paper/2609.08175/) (2026) - [ADIAS: Automated Design of Interactive Agentic Systems](https://mlsanity.com/paper/2608.06410/) (2026) - [Agents in the Large: Perception-Centered Architecture for Persistent Agents](https://mlsanity.com/paper/2608.30478/) (2026) - [Reflections on Trusting Trust, Revisited: Contaminating Self-Modifying AI Coding Agents with Poisoned Benchmarks](https://mlsanity.com/paper/2609.17817/) (2026) - [From Existing Methods to New Research Ideas: Two Perspectives on Recursive Self-Improvement](https://mlsanity.com/paper/2610.14625/) (2026) - [HarnessNet: Feedback-driven Hierarchical Joint Harness Evolution for Multi-agent Systems](https://mlsanity.com/paper/2610.15221/) (2026) - [Self-Evolving Agent Harness](https://mlsanity.com/paper/2610.18813/) (2026) ## Abstract (arXiv) Autonomous agentic systems are largely static after deployment: they do not learn from user interactions, and recurring failures persist until the next human-driven update ships a fix. Self-evolving agents have emerged in response, but all confine evolution to text-mutable artifacts -- skill files, prompt configurations, memory schemas, workflow graphs -- and leave the agent harness untouched. Since routing, hook ordering, state invariants, and dispatch live in code rather than in any text artifact, an entire class of structural failure is physically unreachable from the text layer. We argue that source-level adaptation is a fundamentally more general medium: it is Turing-complete, a strict superset of every text-mutable scope, takes effect deterministically rather than through base-model compliance, and does not erode under long-context drift. We present MOSS, a system that performs self-rewriting at the source level on production agentic substrates. Each evolution is anchored to an automatically curated batch of production-failure evidence and proceeds through a deterministic multi-stage pipeline; code modification is delegated to a pluggable external coding-agent CLI while MOSS retains stage ordering and verdicts. Candidates are verified by replaying the batch against the candidate image in ephemeral trial workers, then promoted via user-consent-gated, in-place container swap with health-probe-gated rollback. On OpenClaw, MOSS lifts a four-task mean grader score from 0.25 to 0.61 in a single cycle without human intervention.