AI ethics evidence index Search
Source-linked daily record
Evidence record 3456 · automatically gathered

COMPASS: Cognitive MCTS-Guided Process Alignment for Safe Search Agents

LLM-powered search agents enable multi-step reasoning and tool use. However, these capabilities introduce retrieval-induced safety degradation, as harmful intents may decompose into seemingly innocuous sub-queries that lead to unsafe outcomes. Existing alignment methods struggle to capture sparse safety signals and fail to supervise diverse violations across multi-step interactions. We propose COMPASS, a Cognitive MCTS-Guided Process Alignment framework designed to achieve robust safety alignmen

Record details

Published: 29 May 2026
Source: arXiv
Category: Research
Topics: Safety & alignment · Agents & autonomy
Retrieved: 14 July 2026

source-onlyevidence status

These records share source-supplied organisations, an exact publisher byline, automatic topics or regions. The reason is shown on every link; related does not mean supporting, agreeing with or verifying this record.

Shared topic: Safety & alignment · Shared topic: Agents & autonomy

TUX: Measuring Human--AI Tacit Understanding

arXiv · 29 May 2026

Shared topic: Safety & alignment · Shared topic: Agents & autonomy

Gram: Assessing sabotage propensities via automated alignment auditing

arXiv · 28 May 2026

Shared topic: Safety & alignment · Shared topic: Agents & autonomy

ROGUE: Misaligned Agent Behavior Arising from Ordinary Computer Use

arXiv · 29 May 2026

Shared topic: Safety & alignment · Shared topic: Agents & autonomy

AgentDoG 1.5: A Lightweight and Scalable Alignment Framework for AI Agent Safety and Security

arXiv · 28 May 2026

Shared topic: Safety & alignment · Shared topic: Agents & autonomy

How Coding Agents Fail Their Users: A Large-Scale Analysis of Developer-Agent Misalignment in 20,574 Real-World Sessions

arXiv · 28 May 2026

Shared topic: Safety & alignment · Shared topic: Agents & autonomy

SPADER: Step-wise Peer Advantage with Diversity-Aware Exploration Rewards for Multi-Answer Question Answering

arXiv · 30 May 2026

How to cite this record

ethics.ai (29 May 2026), “COMPASS: Cognitive MCTS-Guided Process Alignment for Safe Search Agents,” evidence record 3456, https://ethics.ai/record/3456 (originally published by arXiv).

JSON

Use and limitations

This page is a stable index and citation surface for a source record. ethics.ai did not author the underlying report and has not independently verified every claim. Automatic topics may be imperfect. For consequential use, quote and cite the original publisher.