New Method Helps AI Coding Agents Remember and Follow User Corrections Over Time
Researchers have introduced TRACE, a pipeline that converts user corrections to AI coding agents into persistent runtime rules that must be satisfied before task completion. The work addresses a known gap in current memory-based systems, where tools like Mem0 still leave 57.5% of applicable user preferences violated. If validated broadly, the approach could reduce repetitive user frustration and improve the reliability of AI agents in professional workflows.
A team of researchers from multiple institutions has proposed Test-time Rule Acquisition and Compiled Enforcement (TRACE), a skill-layer pipeline designed to make AI coding agents more reliably responsive to user preferences over time. Current memory systems for large language model (LLM) agents often fail to enforce stored preferences: the study found that Mem0, a leading memory tool, still violates 57.5% of applicable preference checks. TRACE addresses this by mining user chat corrections, rewriting them as atomic rules, and compiling those rules into runtime checks that the agent must pass before completing future tasks. In evaluations on ClawArena coding-agent tasks, TRACE reduced held-out preference violations from 100% to 37.6% on in-distribution tasks and from 100% to just 2.0% on out-of-distribution tasks. On MemoryArena-derived memory-intensive tasks, violations dropped from 100% to 60.5%, while task completion performance matched or exceeded the strongest memory baseline. The authors argue this demonstrates that runtime enforcement of compiled corrections addresses a repeated-friction failure mode that memory retrieval alone cannot reliably solve. Both experiment code and a deployable version of the skill are publicly available.
What's missing
The study relies on simulated user-in-the-loop experiments rather than real-world deployment with live users, which may limit generalizability. It is unclear how TRACE handles conflicting or evolving user preferences over time, or how the system performs as the number of compiled rules grows large. The computational overhead of runtime enforcement checks relative to standard memory retrieval is not discussed. The paper has not yet undergone formal peer review, as it is a preprint.
What different sources said
- arXiv cs.CLCenter
Getting Better at Working With You: Compiling User Corrections into Runtime Enforcement for Coding Agents
Related
Gut Bacteria Enzyme Found to Break Down Heat-Processed Food Compounds, Producing Novel Biogenic Amines
Researchers have discovered that an enzyme in common gut bacteria can degrade N-epsilon-carboxymethyllysine (CML), a compound formed during thermal food processing, producing previously unknown biogenic amines. The enzyme, ornithine decarboxylase SpeC from enterobacteria, acts on CML and related modified lysine derivatives through a low-level 'underground' catalytic activity. This finding suggests a previously unrecognized communication axis between thermally processed dietary compounds and gut microbial physiology, with potential implications for host health.
Full-Length Gene Sequencing Reveals Two Distinct Bacterial Communities in Black-Legged Ticks Expanding Into Canada
Researchers used Oxford Nanopore full-length 16S rRNA gene sequencing to characterize the microbiome of Ixodes scapularis black-legged ticks collected in Nova Scotia, Canada, distinguishing between tick-adapted bacteria and environmentally acquired bacteria. The study comes as I. scapularis — the primary vector of Lyme disease — is rapidly expanding northward into Canada due to climate change. The findings suggest that environmentally derived bacteria in tick microbiomes are not mere contamination, which has implications for how tick microbiome data is collected and interpreted across surveillance studies.
Study Identifies Metabolic Link Between Cell Envelope Stress and Biofilm Formation in Bacteria
Researchers have discovered that the metabolite acetyl-CoA directly inhibits enzymes that degrade the bacterial signaling molecule c-di-GMP, connecting cell envelope biosynthesis stress to biofilm formation in Pseudomonas aeruginosa. The study found that sub-inhibitory concentrations of antibiotics targeting early peptidoglycan biosynthesis — but not other antibiotic classes — elevate c-di-GMP levels by reducing phosphodiesterase activity, with acetyl-CoA competing for the enzyme active site. Because the relevant enzyme domain is broadly conserved across bacterial species, this checkpoint mechanism may be widespread and could have implications for understanding antibiotic-induced biofilm responses.