acceptodds
Under review as a conference paper at ICLR 2027

When Consent Outlives Context: Residual Authority Replay in Long-Lived Agents

Abstract

LLM agents increasingly rely on user approval to authorize security-sensitive actions at runtime. Such approvals are granted within a specific task and execution context. In long-lived agents, authorization decisions may need to persist across tasks or sessions. We find that this continuity can outlive the context that originally justified the approval, creating residual authority reusable without renewed consent. We expose this failure mode through a longitudinal attack that starts from a target security-sensitive action, identifies the authority required to execute it, induces benign interactions that legitimately obtain that authority, and later replays the residual authority during adversarial execution. Across controlled and live settings, we demonstrate that residual-authority replay arises in practice and substantially increases the success of prompt-injection and context-rebinding attacks. We evaluate 508 AgentDojo attack cases across six LLM families using production-derived authorization semantics. With residual authority, attack success rate (ASR) increases by up to 35.1 percentage points compared with a fresh authorization state. In live context-rebinding attacks on 55 Terminal-Bench cases across three real-world production coding agents, residual-authority replay increases ASR by 24.9 percentage points on average. These findings expose a fundamental mismatch between persistent authorization and the contextual nature of user consent in long-lived LLM agents.

Primary area:
alignment, fairness, safety, privacy, and societal considerations
Keywords:
LLM agents, agent security, persistent authorization, prompt injection

Reviews

Among:
Shape:
1234this paper, Bottom 10%
Figure 1. Mean review rating of the reviewed papers at ICLR 2027. This paper averages 1.67 over 3 reviews, above 10% of the others.
Reviewer_2ZUJOpenReview ↗
Rating
1
Confidence
4
Informativeness
1
Presentation
3
Soundness
1
Read review
Paper summary

This paper focuses on security risks arising from the persistence of authorization decisions in long-lived LLM agents. The authors introduce the notion of residual-authority replay, in which permissions legitimately granted during benign tasks remain effective and are subsequently exploited in a different, potentially malicious execution context. The paper considers the authorization mechanisms in eight production coding agents and proposes a methodology for identifying permissions required by a target attack, acquiring these permissions through benign interactions, and subsequently reusing them during adversarial execution. The evaluation includes 508 AgentDojo attack cases across six LLM families, as well as 55 Terminal-Bench cases involving three production coding agents. The results show that persistent authorizations can significantly increase the success rate of attacks that would otherwise encounter user-approval requirements.

Critical strengths

The paper addresses a practically relevant security problem in increasingly autonomous and long-lived AI agents. The comparative analysis of authorization mechanisms in existing agent systems is useful, as it highlights differences in permission granularity, persistence, and reuse. The paper also includes the results of a large experimental evaluation. The distinction among harmful-action attempts, authorization decisions, and successful execution helps clarify the specific contribution of retained permissions to attack success.

Critical weaknesses
  1. The security problem addressed by the paper has very limited novelty. My primary concern is the limited conceptual novelty of the paper. The security problem investigated by the authors, that is, that permissions granted for a particular task or context may remain effective after that task or context has ended, is well established in access control and authorization research. Fundamental principles such as least privilege, context-sensitive authorization, limited permission lifetime (see for example the notion of TRBAC), and revocation have long addressed precisely these concerns. Permissions should be constrained not only in terms of the operations and resources they authorize, but also with respect to the conditions under which they may be exercised. The paper tries to present the residual-authority replay as a new attack category. However, the underlying mechanism appears to be a specific manifestation of a known authorization problem: permissions that are insufficiently restricted in scope or lifetime can subsequently be reused in unintended contexts. The paper acknowledges closely related research on authorization continuity, lingering authority, and task-scoped authorization. However, I'm not convinced that the distinction drawn between this prior research and residual-authority replay establishes a sufficiently substantial conceptual contribution.

  2. The proposed attack methodology has limited technical novelty. The proposed methodology consists of identifying the permissions necessary for a target malicious operation, finding benign tasks through which those permissions can be legitimately acquired, and subsequently exploiting the accumulated permissions in a different execution context. Although the methodology is clearly presented, its technical novelty is not apparent. The backchaining process essentially maps required operations to previously acquired authorization grants. Given an authorization system in which permissions persist across tasks, it is reasonably foreseeable that an attacker could exploit benign interactions to accumulate the permissions necessary for subsequent malicious operations. Actually in real life attacks, one step often carried out by the attackers is to escalate privileges in various ways (see for example the well known 2014 Target attack, in which the attackers were able to stole a token used by a system administrator for single-sign on and then used the privileges of the administrator, to create their own system administrator account). So the attack approach proposed in the paper is one of the possible ways agents could get permissions. It is therefore unclear whether the proposed approach introduces a genuinely new attack-generation technique or simply provides a systematic experimental procedure for demonstrating the consequences of persistent permissions.

  3. The authorization problems in agentic systems is narrowly addressed. Another important limitation is that the paper investigates only a relatively narrow dimension of authorization security in agentic systems. Authorization in agentic environments involves substantially more than the persistence of permissions across tasks. In particular, multi-agent systems introduce additional complexities involving permission delegation, transitive authorization, accumulation of privileges, and preservation of authorization constraints across agent interactions. For example, an agent may receive permissions for a particular task and subsequently delegate some of these permissions to other agents. Unless delegation is carefully constrained, downstream agents may acquire permissions without the original task scope, execution context, or delegation restrictions being preserved. Moreover, permissions obtained through different delegation chains or accumulated over multiple tasks may collectively provide an agent, or a cooperating group of agents, with significantly more authority than intended. These issues are closely related to the problem addressed by the paper but are not considered in its threat model or evaluation. I am not suggesting that the paper must address all aspects of multi-agent authorization. However, the authors should clarify that their contribution is narrow and limited, and avoid presenting permission persistence as a newly identified or comprehensive authorization problem for long-lived agents.

  4. The experimental results largely confirm expected consequences of persistent permissions and are thus not surprising. The experimental evaluation is extensive, but the significance of its findings should be considered in relation to the novelty of the underlying problem. The experiments compare attacks under fresh authorization states with attacks under states containing previously acquired permissions. In the controlled evaluation, unauthorized security-sensitive operations are stopped at the approval boundary in the fresh state. When previously acquired permissions match those operations, the operations can execute without renewed approval. Consequently, the observed increases in attack success rates are largely consistent with the expected behavior of the authorization mechanisms being evaluated. The experiments quantify the magnitude of this effect, which is useful. However, quantitative confirmation of an established security weakness does not, by itself, establish a substantial new technical contribution. Furthermore, the evaluation of task-aware authorization mechanisms demonstrates that restricting permissions to the current task can substantially reduce or eliminate the tested replay attacks. This is consistent with established least-privilege and context-sensitive authorization principles.

Questions for authors

None

Reviewer_VyexOpenReview ↗
Rating
2
Confidence
3
Informativeness
1
Presentation
2
Soundness
2
Read review
Paper summary

This paper examines how approvals granted during benign tasks can remain usable after the execution context changes. It analyzes authorization mechanisms in production coding agents and proposes a process for acquiring and replaying retained permissions. Experiments and live-agent tests show that these permissions can enable malicious actions without renewed consent.

Critical strengths
  1. The source-level analysis connects the security risk to concrete authorization choices in production agents.
  2. Separating harmful-action attempts from authorization and execution outcomes helps explain why retained permissions amplify attacks.
Critical weaknesses
  1. The effectiveness of deliberate authority acquisition remains unclear. The experiments use untargeted histories, while acquisition success, failures, and approval costs are not reported across the live cases.
  2. The live results are difficult to verify against the stated replay criterion. Success rate gains need clearer supporting records showing that retained grants admitted the corresponding actions without renewed approval.
Questions for authors

How reliably can an attacker acquire the required permissions through benign tasks, and how much interaction and approval effort does this require?

Reviewer_A8DLOpenReview ↗
Rating
2
Confidence
3
Informativeness
2
Presentation
2
Soundness
2
Read review
Paper summary

This paper investigates residual-authority replay in long-lived LLM agents: permissions acquired during legitimate interactions remain effective even after the context that justified them has changed. The paper also examines how the length of authorization history and permission granularity affect this amplification.

Critical strengths
  • The paper provides useful evidence about authorization persistence in production agents.
  • The history-length sweep and authorization-granularity comparisons provide evidence that exposure depends on both accumulated authority and the specificity of permission matching.
Critical weaknesses
  • Regarding persistence, Table 2 varies the number of preceding benign tasks, (h), while the nested histories monotonically accumulate grants. This establishes an effect of accumulated permission coverage, but does not independently characterize permission age, retention windows, or the interval between acquisition and exploitation.
  • The source-level analysis checks persistence across execution boundaries, but the quantitative evaluation does not separate these lifecycle effects from the increasing number and diversity of grants. In particular, Goose’s exposure saturates after one task in Figure 2, suggesting that permission granularity and coverage can dominate history length. The paper should clarify which temporal properties are actually measured and characterize attack effectiveness under bounded retention or revocation.
  • The high-level principle that authorization should depend on the current task and context is already addressed by related work. Lingering Authority directly studies temporary authority that survives its justifying episode and rejects stale reuse through capability revocation; Sudobench also studies the related question. The current comparisons do not resolve this question. Appendix D implements only PAuth’s task-scoping component and disables Progent’s dynamic policy updates.

[1] Lingering Authority: https://arxiv.org/pdf/2606.22504 [2] SudoBench: A Contextual Authorization Benchmark for LLM Agents

Questions for authors
  • What precisely does “length” represent: prior tasks, distinct grants, or retained authorization entries?
  • For each evaluated agent, what terminates a grant’s validity? Can you separate grant accumulation from retention effects by holding permission coverage fixed while varying intervening tasks and lifecycle boundaries, and report sensitivity to bounded retention or revocation?

From OpenReview, CC BY 4.0. Authors as signed there; shown unedited.

Related papers

Loading the map…

Discussion (0)

Sign in to comment.