Agentic Harnesses: LLM-Driven Verification Layers for Robot Autonomy

Source

• Source name: arXiv
• Source kind: paper
• Authors / organization: Rohan Bhagra, Mahantesh Halapannavar, Uddhav Bhattarai
• Venue: arXiv: cs.RO
• Published: 2026-08-10
• Collected: 2026-08-11
• URL: http://arxiv.org/abs/2608.09857v1
• PDF: https://arxiv.org/pdf/2608.09857v1
• DOI: UNKNOWN
• Registry group: academic_apis_and_preprints
• Evidence role: canonical_academic_index
• Language: en
• Country / region: UNKNOWN

Plain-Language Brief

The authors propose a system that uses large language models (LLMs) to evaluate the permissibility of actions planned by robots before they are executed. The system aims to reduce risks associated with biased planning models, unsafe actions, and vulnerabilities to adversarial attacks.

What the Document Says

• The paper discusses a verification layer for robot autonomy that uses LLMs.
• The verification layer is designed to evaluate the feasibility of actions proposed by robots' planning models.
• The system is aimed at addressing risks associated with biased and unsafe actions, as well as vulnerabilities to adversarial attacks in robotics planning models.
• The LLM-as-a-Judge ensemble uses chain-of-thought reasoning across multiple models and synthesizes the outputs.
• This system serves as middleware, filtering plans from the server's planning module before they reach the MCP server and the robot's low-level controls.
• The system achieves near 85% precision in accepting/escalating/rejecting tasks, with a containment of 97% for adversarial attacks.
• Errors occur mostly at the escalate boundary, with negligible errors between accepting and rejecting tasks.

Key Concepts Mentioned

• robot autonomy
• large language models (LLMs)
• action permissibility
• verification layer
• chain-of-thought reasoning
• adversarial attacks

Claims Made by the Document

• The proposed system reduces risks in robotics planning models.
• The system achieves near 85% precision and 97% containment of adversarial attacks.

Evidence Used

• The text describes the system's purpose and its performance metrics.

Methods / Approach

• Chain-of-thought reasoning across multiple LLMs
• Synthesizing outputs from expert judges

Limitations Stated by the Authors

UNKNOWN

Limitations Noted by the Agent

Agent interpretation — not from the source.

UNKNOWN

Open Questions

UNKNOWN

Notes for Future Q&A

May include interpretation. Labeled where it does.

UNKNOWN

Raw Abstract / Excerpt

Advances in advanced artificial intelligence tools have sparked research in robot autonomy, but the development of such systems has largely focused on execution rather than verifying the feasibility actions planning models propose. Like general-purpose LLMs, robotics planning models carry risks: biased toward user-specified goals, they may suggest actions misaligned with scientific ethics, they may be unsafe due to an inability to 'remember' prior safety risks, or they may be vulnerable to adversarial attacks on the autonomy ecosystem. We propose a LLM-driven verification layer between planning and execution to evaluate action permissibility. Our LLM-as-a-Judge ensemble combines chain-of-thought reasoning across models and synthesizes those expert judge outputs, mirroring a combination of a mixture of experts and self-consistency approach. This layer serves as middleware, gating plans from the server's planning module before they reach the MCP server and therefore the robot's low-level controls: plans are approved, rejected for reformulation, or escalated for human review. With this system, we achieve near 85% precision across accept/escalate/reject categories 97% containment of adversarial attacks, with negligible errors between accepting and rejecting tasks, and errors mostly manifesting at the escalate boundary.