Xiaochong Jiang

Su Wang, Pin Qian, Yifan Lin et al.

Self-improving AI agents are designed to learn from their mistakes. We show they can also hallucinate mistakes that never happened. We study this failure mode in automated harness optimization, where an LLM-based proposer edits an agent's scaffold, including prompts, parsers, filters, validators and guardrails, to eliminate observed failures. But this process rarely asks first: was there a real failure to fix? We introduce the Counterfactual Fabrication Lab, a deterministic micro-lab where the correct action is known: do nothing. The lab plants a candidate guardrail for a failure class that provably never occurs, presents only legal episodes, and uses a byte-exact oracle to check every cited violation. The proposer behaves as expected on real violations and abstains on featureless legal input. Yet when the legal input contains a harmless pattern resembling a familiar game rule, it invents a failure: in 15/60 runs, versus 0/60 on featureless input, it enables the nonexistent-rule guardrail and cites a violation the oracle refutes. The effect is structured, not indiscriminate. In single-shot proposals it appears only when three conditions coincide: a rule-shaped pattern, an open-ended rule set and an instruction that presupposes failures. Removing any of these conditions eliminates the fabrication. Because the invented guardrail changes no true outcome and cannot improve an already-perfect suppression score, the phenomenon is neither reward hacking nor over-refusal. It is a phantom guardrail: a fix for a failure that never happened, invisible to suppression-only acceptance. Inside an add-only accept loop it re-enters even without the failure-presupposing instruction, the loop's keep-adding role supplying the demand the instruction supplied in single shot, and once in it stays. We present the Counterfactual Fabrication Lab for measuring fabricated failures in self-improving agent harnesses.

Su Wang, Pin Qian, Yihang Chen et al.

LLM agents increasingly rely on community-contributed skills that expand an agent's operational capability set. We study a core safety problem in agentic AI systems: whether individually safe skills can compose into unsafe installed skill sets. We present SkillReact, a compositional security measurement framework with three components: a deterministic static-composition benchmark, a two-rater LLM-assisted human-adjudication pipeline, and an action-based exploitability harness. On 1,520 ClawHub skills, 651 pass individual inspection and form 211,575 pairs; the benchmark flags 22.25% of these as structural candidates. We treat this raw rate as a recall-oriented scanner ceiling and calibrate it against human judgment: in a pattern-stratified audit, roughly one in five flagged pair-pattern hits survives as a real compositional risk (population-weighted validity 18.2%, our headline result), implying about 14K genuine risk memberships in a single registry that per-skill scanning misses by construction, since every pair is individually safe. An action-based harness then probes when these candidates become model-issued tool calls, and finds realization gated by host-model disposition: on an anchor-conditioned dropper subset, Haiku-4-5 issues the dropper-stage tool call on all 39 direct-prompt trials (36 of them the full download-then-execute chain, 3 download-only), Opus-4-7 stops at the download, and Sonnet-4-6 refuses outright. A control that holds the request fixed and varies only the installed skills finds compliance highest with no skills installed: a composition fixes which capabilities are reachable, while the host model decides whether to use them. Together these motivate install-time compositional checks and capability isolation as complements to per-skill scanning.

Lifei Liu, Haoran Yu, Xiaochong Jiang et al.

Safety evaluations of multi-agent LLM systems often compare a direct prompt with a planner-executor pipeline and report the difference as a single "pipeline effect." We argue that this aggregate is difficult to interpret because it conflates three mechanisms: harmful intent may be reframed as plausible operational work, the planner may refuse or transform the request, and the executor may act under delegation prompts implying prior approval. To separate these factors, we introduce a five-condition controlled contrast design, evaluated on 30 synthetic harmful scenarios and an exploratory external validation set from four agent-safety benchmarks using LLM-judged compliance. Our results show that aggregate pipeline safety is not a stable architectural property. Operational reframing is the most portable risk signal, increasing compliance for GPT, Gemini, and DeepSeek across both scenario sets, while Claude is comparatively resistant. Planner behavior can offset this risk mainly through refusal; however, when the planner produces executable steps, the executor may become more compliant than under the direct operational baseline. Approval-framed delegation is sensitive to prompt design, model pairing, and scenario source, and a skeptical executor prompt sharply reduces compliance. Raw-direct model rankings can also mispredict deployed planner-executor behavior. Gemini is safest under raw direct prompts in the primary set yet shows the largest amplification with a Claude planner, rising from 8.9 percent to 38.9 percent compliance. GPTs near-zero aggregate pipeline effect instead hides a reframing increase canceled by planner refusal. These findings suggest that multi-agent safety evaluations should report reframing, planner behavior, delegation framing, and model pairing separately before attributing failures to architecture itself.

Xiaochong Jiang, Shiqi Yang, Ziwei Li et al.

Tool-using agents increasingly operate in open-ended deployment environments, where they compose file systems, web APIs, code interpreters, and enterprise services at runtime. This creates a safety gap in tool composition: an agent can satisfy every per-tool permission check and still produce an unsafe end-to-end effect, such as reading a confidential document, summarizing it, and sending the summary to an external endpoint. We call this failure mode permission laundering. ChainCaps addresses it with a runtime rule: every value carries a sink-specific capability budget, and tool composition propagates budgets by intersection. A value can preserve or lose authority as it moves through a tool chain, but it cannot gain new authority through composition. We implement ChainCaps as a transparent MCP proxy that requires no changes to the agent or tool servers. On 82 tasks across five frontier models from three providers, ChainCaps reduces attack success rate from 25-68% to 0-4.8% while preserving 96-100% benign completion. In replay experiments, it also outperforms scalar-IFC and per-function-isolation baselines. Manifest quality is the dominant deployment bottleneck: expert manifests reach 100% attack blocking, while naive manifests fall to 27.3%. Our claims are limited to explicit-flow composition safety under trusted manifests and proxy-visible data movement, a practical gap in deployed tool-using agents today.

Yichen Liu, Imam Akintomiwa Akinlade, Xiaochong Jiang et al.

Environmental monitoring is a crucial component of the smart city infrastructure. It enables informed decision making which enhances sustainability, public health and urban planning. However, the large-scale deployments of the smart sensors have raised concerns on excessive energy consumption and redundant data collection as well as limited sensor lifespan. To resolve these issues, we present an AI-driven framework for energy-efficient environmental monitoring in smart cities utilizing edge intelligence. Our proposed framework leverages TinyML-enabled edge devices and context-aware adaptive decision-making in order to dynamically activate the sensors based on the spatiotemporal conditions, environmental statistics and energy constraints. The sensors will be dynamically activated based on a utility function that takes in factors such as real-time environmental conditions, sensor location, and remaining battery lifespan. Our framework will reduce unnecessary sensing and communication while maintaining high coverage for monitoring. We introduce a hierarchical Edge Intelligence architecture to support deployments in city-wide scales. We conducted evaluation using a city-scale simulation driven by real multi-sensor environmental traces, which demonstrates that the proposed mechanism significantly reduces energy consumption and extends sensor lifespan when compared to static, periodic, and UCB-based adaptive sensing strategies. The results highlight the potential of edge intelligence and adaptive AI techniques for building sustainable and efficient smart city monitoring systems.

Xiaochong Jiang, Shiqi Yang, Wenting Yang et al.

Agentic systems built on large language models (LLMs) extend beyond text generation to autonomously retrieve information and invoke tools. This runtime execution model shifts the attack surface from build-time artifacts to inference-time dependencies, exposing agents to manipulation through untrusted data and probabilistic capability resolution. While prior work has focused on model-level vulnerabilities, security risks emerging from cyclic and interdependent runtime behavior remain fragmented. We systematize these risks within a unified runtime framework, categorizing threats into data supply chain attacks (transient context injection and persistent memory poisoning) and tool supply chain attacks (discovery, implementation, and invocation). We further identify the Viral Agent Loop, in which agents act as vectors for self-propagating generative worms without exploiting code-level flaws. Finally, we advocate a Zero-Trust Runtime Architecture that treats context as untrusted control flow and constrains tool execution through cryptographic provenance rather than semantic inference.