ParanoiaEval: Benchmarking Unnecessary Defensive Work in Agentic Coding
cs.AI, cs.SE
Submitted: 2026-10-06
Updated: 2026-10-15
Code: https://github.com/ZhuoningXu/ParanoiaEval_release
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Agentic Coding Needs Proactivity, Not Just Autonomy
- Coding Agents Don't Know When to Act
- SWE-Factory: Your Automated Factory for Issue Resolution Training Data and Evaluation Benchmarks
- Redundant or Necessary? A Benchmark for Detecting Redundant Steps in Agent Trajectories
- Professional Software Developers Don't Vibe, They Control: AI Agent Use for Coding in 2025
- From LLMs to LLM-based Agents for Software Engineering: A Survey of Current, Challenges and Future
- Configurable AI Coding Assistants: Designing For Developers Who Like to Be in Control
- Large Language Model-Based Agents for Software Engineering: A Survey
- FAIntbench: A Holistic and Precise Benchmark for Bias Evaluation in Text-to-Image Models
- BIGbench: A Unified Benchmark for Evaluating Multi-dimensional Social Biases in Text-to-Image Models
- AgentAuditor: Human-Level Safety and Security Evaluation for LLM Agents
- AtelierEval: Agentic Evaluation of Humans & LLMs as Text-to-Image Prompters
- BiasIG: Benchmarking Multi-dimensional Social Biases in Text-to-Image Models
- AutoCRAT: Within-trajectory Joint Control of Stochasticity and Compute for LLM Reasoning
- CentaurEval: Benchmarking Human-in-the-Loop Value in Agentic Coding
- Beyond Resolution Rates: Behavioral Drivers of Coding Agent Success and Failure
- Terminal-Bench: Benchmarking Agents on Hard, Realistic Tasks in Command Line Interfaces
- Wink: Recovering from Misbehaviors in Coding Agents
- Nomic Embed: Training a Reproducible Long Context Text Embedder
- SNARE: Adaptive Scenario Synthesis for Eliciting Overeager Behavior in Coding Agents
Related papers
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Agents' Last Exam
- The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
- MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
- Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems