# Understanding and Exploiting Initialization Anchoring Weakness in Feedback-Based Agent Planning

arXiv: https://arxiv.org/abs/2609.29697 · HTML: https://mlsanity.com/paper/2609.29697/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Agent threats, guardrails and governance, one of 16 directions in LLM agents and tool use.
- Builds on 20 of its 61 references that are in the map.
- First posted on arXiv 31 Aug 2026.

## Placed in

- LLM agents and tool use / [Agent threats, guardrails and governance](https://mlsanity.com/agents/agent-threats-guardrails-and-governance/)

## Builds on (most cited first, 20 of 20)

- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIs](https://mlsanity.com/paper/2307.16789/) (2023)
- [AgentBench: Evaluating LLMs as Agents](https://mlsanity.com/paper/2308.03688/) (2023)
- [Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations](https://mlsanity.com/paper/2312.06674/) (2023)
- [Ignore Previous Prompt: Attack Techniques For Language Models](https://mlsanity.com/paper/2211.09527/) (2022)
- [GPTFUZZER: Red Teaming Large Language Models with Auto-Generated Jailbreak Prompts](https://mlsanity.com/paper/2309.10253/) (2023)
- [SmoothLLM: Defending Large Language Models Against Jailbreaking Attacks](https://mlsanity.com/paper/2310.03684/) (2023)
- [Detecting Language Model Attacks with Perplexity](https://mlsanity.com/paper/2308.14132/) (2023)
- [Understanding the planning of LLM agents: A survey](https://mlsanity.com/paper/2402.02716/) (2024)
- [Defeating Prompt Injections by Design](https://mlsanity.com/paper/2503.18813/) (2025)
- [WASP: Benchmarking Web Agent Security Against Prompt Injection Attacks](https://mlsanity.com/paper/2504.18575/) (2025)
- [Thought Anchors: Which LLM Reasoning Steps Matter?](https://mlsanity.com/paper/2506.19143/) (2025)
- [ReST meets ReAct: Self-Improvement for Multi-Step Reasoning LLM Agent](https://mlsanity.com/paper/2312.10003/) (2023)
- [MELON: Provable Defense Against Indirect Prompt Injection Attacks in AI Agents](https://mlsanity.com/paper/2502.05174/) (2025)
- [UDora: A Unified Red Teaming Framework against LLM Agents by Dynamically Hijacking Their Own Reasoning](https://mlsanity.com/paper/2503.01908/) (2025)
- [ObliInjection: Order-Oblivious Prompt Injection Attack to LLM Agents with Multi-source Data](https://mlsanity.com/paper/2512.09321/) (2025)
- [Cognitive Overload Attack:Prompt Injection for Long Context](https://mlsanity.com/paper/2410.11272/) (2024)
- [From Risk Classification to Action Plan Remediation: A Guardrail Feedback Driven Framework for LLM Agents](https://mlsanity.com/paper/2606.05805/) (2026)
- [Measuring Real-World Prompt Injection Attacks in LLM-based Resume Screening](https://mlsanity.com/paper/2605.28999/) (2026)
- [PlanGuard: Defending Agents against Indirect Prompt Injection via Planning-based Consistency Verification](https://mlsanity.com/paper/2604.10134/) (2026)

## Abstract (arXiv)

Feedback-based planning improves agent reliability by incorporating tool observations and corrective feedback. However, its protection may not be distributed uniformly across planning stages. We conduct a round-wise analysis of four representative feedback mechanisms and uncover an initialization anchoring weakness: the first feedback round corrects 46% of adversarial directions, whereas the rates fall to 13% and 7% among directions surviving into the next two rounds. Our analysis attributes this weakness to three interacting factors: a contextually plausible shift in the initial plan, insufficient counterevidence, and the persistence of accepted directions in the accumulated trajectory. Based on these findings, we propose textscInitAnchor, a black-box framework for exploiting this weakness through attacker-controlled external materials. It operationalizes the three factors as directional-shift, contextual-plausibility, and counterevidence-resilience signals under either limited target access or no target access. Across 112 tasks from 16 domains, six agent architectures, and five backbone LLMs, textscInitAnchor achieves average ASRs of 76.1% and 72.0% under the two settings while reducing first-round mitigation rates to 21.0% and 25.0%, respectively. It also remains effective against six defenses and across six real-world agent systems. These findings show that feedback-based agents can retain early biases even when later correction is available.