Ratchet: A Minimal Hygiene Recipe for Self-Evolving LLM Agents

• URL: http://arxiv.org/abs/2605.22148v1
• Date: Thu, 21 May 2026 08:20:38 GMT
• Title: Ratchet: A Minimal Hygiene Recipe for Self-Evolving LLM Agents
• Abstract summary: We introduce textbfRatchet, a single-agent loop in which a frozen LLM writes, retrieves, curates, and retires its own natural-language skills.<n>Ratchet integrates four candidate hygiene mechanisms: outcome-driven retirement, a bounded active-cap, meta-skill authoring guidance, and pattern canonicalisation.
• Score: 9.989306175511238
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Self-evolving skill libraries, pioneered by Voyager, let frozen LLM agents accumulate reusable knowledge without weight updates, yet recent evaluation shows that LLM-authored skills deliver $+0.0$pp over no-skill baselines while human-curated ones deliver $+16.2$pp: the bottleneck is not skill authoring but lifecycle management. We introduce \textbf{Ratchet}, a single-agent loop in which a frozen LLM writes, retrieves, curates, and retires its own natural-language skills. Ratchet integrates four candidate hygiene mechanisms: outcome-driven retirement, a bounded active-cap, meta-skill authoring guidance, and pattern canonicalisation. On MBPP+ hard-100 with Claude Opus 4.7, Ratchet lifts held-out pass@1 from a $0.258 \pm 0.047$ baseline to a late-window rolling mean of $0.584$ (peak $0.658 \pm 0.042$) across 100 rounds and 3 seeds, a $+0.328 \pm 0.018$ rolling-mean gain where the no-skill control drifts at $+0.002 \pm 0.005$; the same recipe transfers to an agentic solver on SWE-bench Verified ($+0.22$ peak lift over 20 rounds). Eight ablations (A1--A8) reveal that the minimal working recipe is smaller than our design suggests: retirement and the meta-skill authoring prior are load-bearing, while explicit deduplication (canonicalisation, cover-guard) is subsumed by the meta-skill itself. A non-divergence proposition shows that bounded cap and retirement threshold together prevent expected performance from drifting below the no-skills floor.

Related papers

• SANE: State Anomaly Neutralization for Stable Extreme-Context Delta-Rule Models [53.86918766240095]
  Delta-Rule recurrent models maintain a fixed-size state, enabling $O(1)$ inference memory but potentially becoming unstable under extreme-context extrapolation.<n>We propose textbfState Anomaly Neutralization (SANE), which applies adaptive $tanh$ compression at chunk boundaries while preserving the intra-chunk parallel structure.
  arXiv  Detail & Related papers  (2026-08-23T10:41:07Z)
• ShortOPD: Recovering Pruned LLMs with Short-to-Long On-Policy Distillation [72.54604107217669]
  textscpass@$k$ nearly vanishes after compression, yet textscpass@$k$ recovers substantially under repeated sampling.<n>Recovery should train on the compressed model's own on-policy states with dense token-level supervision.<n>We propose textbfshortopd, a short-to-long OPD schedule that detects teacher-confirmed repetitive suffixes.
  arXiv  Detail & Related papers  (2026-07-14T17:50:50Z)
• Scaling Self-Evolving Agents via Parametric Memory [69.96398842169002]
  Existing memory-augmented LLM agents store past experience exclusively in prompt space.<n>We introduce textttTMEM, a self-evolving parametric memory framework.<n>textttTMEM consistently outperforms summary-based and retrieval-based baselines across different model scales.
  arXiv  Detail & Related papers  (2026-06-03T07:18:31Z)
• Which Defense Closes Which Threat? Attributing OWASP-LLM-Top-10 Coverage and Its Brittleness Under Paraphrasing [51.56484100374058]
  Production LLM applications stack several defense families -- refusal-phrase filters, tokenbudget controls, model allowlists, rate limits, tool-registry authentication -- yetBAS benchmarks report a single aggregate coverage number.<n>We add four-LLM-Top-10-aware agents to a 21-agent baseline scanner and target a lattice of four synthetic LLM endpoints.
  arXiv  Detail & Related papers  (2026-06-01T19:39:25Z)
• When Skills Don't Help: A Negative Result on Procedural Knowledge for Tool-Grounded Agents in Offensive Cybersecurity [1.3591163361204688]
  Skills are packages of procedural knowledge loaded into an agent at inference time.<n>They are widely reported to improve task pass rates by an average of 16.2percentage points across diverse domains.<n>Yet the same benchmarks show wide variance, with 16 of 84 tasks suffering negative deltas when Skills are introduced.
  arXiv  Detail & Related papers  (2026-05-19T15:48:35Z)
• Library Drift: Diagnosing and Fixing a Silent Failure Mode in Self-Evolving LLM Skill Libraries [9.989306175511238]
  Self-evolving skill libraries face a silent failure mode we term emphlibrary drift: skill accumulation without outcome-driven lifecycle management causes retrieval degradation, false-positive injections, and performance stagnation.<n>Recent evaluation confirms the symptom--LLM-authored skills deliver +0.0pp gain while human-curated ones deliver +16.2pp (Bench)
  arXiv  Detail & Related papers  (2026-05-19T09:19:56Z)
• Representation Without Reward: A JEPA Audit for LLM Fine-Tuning [1.2691047660244335]
  Joint-embedding predictive architectures (JEPAs) propose that a model should learn more useful abstractions when trained to predict latent representations rather than observed outputs.<n>For autoregressive language-model fine-tuning the principle entails a stricter requirement: the induced hidden-state geometry must reach the language-model head emphand improve the decoded task metric.<n>We test that requirement under a fixed Llama-3.2-1B-Instruct LoRA harness on natural-language-to-regex generation.
  arXiv  Detail & Related papers  (2026-05-14T20:27:32Z)
• Attributing Emergence in Million-Agent Systems [68.53670424791751]
  Large language models (LLMs) can simulate human-like reasoning and decision-making in individual agents.<n>Such studies require attributing macro emergence to individual agents.<n>We adapt Aumann--Shapley path-integral attribution to LLM-powered MAS at million-agent scale.
  arXiv  Detail & Related papers  (2026-05-12T01:49:41Z)
• SkillEvolver: Skill Learning as a Meta-Skill [46.52001686198623]
  We propose textbfSkillEvolver, a lightweight, plug-and-play solution for online skill learning.<n>The learning target of SkillEvolver is the skill's prose and code, not model weights.<n>Unlike trace-distillation, the meta-skill refines only after deploying the learnt skill.
  arXiv  Detail & Related papers  (2026-05-11T12:58:25Z)
• A Preliminary Agentic Framework for Matrix Deflation [2.6140509675507384]
  We propose an agentic approach to matrix deflation in which a solver generates rank-1 updates and decides when to stop.<n>We evaluate on Digits ($8times8$), CIFAR-10 ($32times32$ grayscale), and synthetic ($16times16$) matrices with and without Gaussian noise.<n>Across all settings, our agent achieves competitive results, suggesting that fully agentic, threshold-free deflation is a viable alternative to classical numerical algorithms.
  arXiv  Detail & Related papers  (2026-01-06T23:59:18Z)
• Data-Free Pruning of Self-Attention Layers in LLMs [1.7188280334580195]
  We propose Gate-Norm, a one-shot, weight-only criterion that ranks attention sublayers by query-key coupling.<n>Gate-Norm removes the least coupled ones, requiring no calibration data, no forward passes, no fine-tuning, and no specialized kernels.
  arXiv  Detail & Related papers  (2025-12-03T07:47:49Z)
• ResT: Reshaping Token-Level Policy Gradients for Tool-Use Large Language Models [62.82372407840088]
  Large language models (LLMs) transcend passive generation and act as goal-directed agents by invoking external tools.<n>textbfReshaped textbfToken-level policy gradients (textbfResT) for tool-use tasks.<n>textbfResT achieves state-of-the-art results, outperforming prior methods by up to $8.76%$.
  arXiv  Detail & Related papers  (2025-09-26T03:38:27Z)
• FedSVD: Adaptive Orthogonalization for Private Federated Learning with LoRA [68.44043212834204]
  Low-Rank Adaptation (LoRA) is widely used for efficient fine-tuning of language models in learning (FL)<n>Low-Rank Adaptation (LoRA) is widely used for efficient fine-tuning of language models in learning (FL)
  arXiv  Detail & Related papers  (2025-05-19T07:32:56Z)
• Skill-Mix: a Flexible and Expandable Family of Evaluations for AI models [50.11814354654953]
  Key ability of an AI agent is to flexibly combine, as needed, the basic skills it has learned. This work introduces Skill-Mix, a new evaluation to measure ability to combine skills.
  arXiv  Detail & Related papers  (2023-10-26T16:55:05Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.