When Valid Tool Calls Change Meaning: Formation-Consistent Dispatch for LLM Agents
On the map: LLM agents and tool use → Software engineering coding agents also fits: Agent threats, guardrails and governance
Abstract (arXiv)
Tool-enabled agents form calls from model-visible interfaces, while hosts later select their implementation. Standard dispatch omits the descriptor-handler relation. An unchanged and schema-valid call can therefore acquire a different security effect during rollout, reconnect, or delayed approval. We call this failure schema-epoch drift. We present formation-consistent dispatch (FCD), which connects implementation analysis to execution authority. Reviewed profiles produce provenance-bound over-approximations of declared in-scope effects from official source. Under a closed-target approval policy, a verifier applies each formed call to a summary and captures a successor only when its effects fit the call's security contract. Atomic admission and a final-hop fence preserve this decision to the effect. The exact source retains priority, and the captured successor becomes eligible only after source retirement. Stock releases and deployment changes reproduced the failure. Four profiles covered 32 official releases: 29 required no release-specific change and three escalated. A frozen 16-release expansion matched a separate source oracle. In a preregistered stock comparison, FCD completed all three pending calls whose effect remained private and blocked all three whose omission became public. Exact pinning and release-wide denial stopped all six calls, while release-wide approval completed all six but produced three public effects. A separate lifecycle experiment carried a formation-captured certificate across source retirement. The same safe certificate installed later governed new formations without expanding the pending call's authority.
Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Software engineering coding agents, one of 16 directions in LLM agents and tool use. Builds on 23 of its 65 references that are in the map. First posted on arXiv 28 Sep 2026.
Builds on
23 of 65 references.
References · 23
- 2023Identifying the Risks of LM Agents with an LM-Emulated Sandboxcited by 514
- 2024InjecAgent: Benchmarking Indirect Prompt Injections in Tool-Integrated Large Language Model Agentscited by 476
- 2025Defeating Prompt Injections by Designcited by 256
- 2025AgentSpec: Customizable Runtime Enforcement for Safe and Reliable LLM Agentscited by 196
- 2025Progent: Securing AI Agents with Privilege Controlcited by 121
- 2024IsolateGPT: An Execution Isolation Architecture for LLM-Based Agentic Systemscited by 118
- 2025Contextual Agent Security: A Policy for Every Purposecited by 55
- 2025ACE: A Security Architecture for LLM-Integrated App Systemscited by 52
- 2025MCPSecBench: A Systematic Security Benchmark and Playground for Testing Model Context Protocolscited by 47
- 2025SAGA: A Security Architecture for Governing AI Agentic Systemscited by 45
- 2026AttriGuard: Defeating Indirect Prompt Injection in LLM Agents via Causal Attribution of Tool Invocationscited by 33
- 2025Towards Automating Data Access Permissions in AI Agentscited by 20
11 more
- 2026Cordon: Semantic Transactions for Tool-Using LLM Agentscited by 19
- 2025ETDI: Mitigating Tool Squatting and Rug Pull Attacks in Model Context Protocol (MCP) by using OAuth-Enhanced Tool Definitions and Policy-Based Access Controlcited by 18
- 2025Parasites in the Toolchain: A Large-Scale Analysis of Attacks on the MCP Ecosystemcited by 15
- 2025Semantic Attacks on Tool-Augmented LLMs: Securing the Model Context Protocol Against Descriptor-Level Manipulationcited by 12
- 2026ChainCaps: Composition-Safe Tool-Using Agents via Monotonic Capability Attenuationcited by 11
- 2025When AI Meets the Web: Prompt Injection Risks in Third-Party AI Chatbot Pluginscited by 6
- 2026CAVA: Canonical Action Verification and Attestation for Runtime Governance of Agentic AI Systemscited by 3
- 2025It's a Feature, Not a Bug: Secure and Auditable State Rollback for Confidential Cloud Applicationscited by 1
- 2026Attested Tool-Server Admission: A Security Extension to the Model Context Protocolcited by 1
- 2026Tool Forge: A Validation-Carrying Toolchain for Governed Agentic Executioncited by 1
- 2026From Tool Connection to Execution Control: Benchmarking Security Invariants in MCP-Style Agent Runtimescited by 1
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2026Tool-Making and Self-Evolving LLM Agents in Low-Latency Systems
- 2026Do Tool Calls Execute as Intended? Measuring and Repairing Intent-Execution Correspondence in LLM Agents
- 2026Verified Tool Calls Improve LLM Agent Reliability Under Non-Atomic Failures
- 2026Temporary Authority, Permanent Effects: Commit-Time Authorization for LLM Agents
- 2026Skill Drift Is Contract Violation: Proactive Maintenance for LLM Agent Skill Libraries
- 2025More Vulnerable than You Think: On the Stability of Tool-Integrated LLM Agents
- 2026Reason Less, Verify More: Deterministic Gates Recover a Silent Policy-Violation Failure Mode in Tool-Using LLM Agents
- 2026Beyond Approved Actions: Runtime Validation of Persistent Outcomes in Agent Workflows