SDOF: Taming the Alignment Tax in Multi-Agent Orchestration with State-Constrained Dispatch

• URL: http://arxiv.org/abs/2605.15204v1
• Date: Mon, 20 Apr 2026 12:51:39 GMT
• Title: SDOF: Taming the Alignment Tax in Multi-Agent Orchestration with State-Constrained Dispatch
• Abstract summary: We present SDOF, a framework that treats multi-agent execution as a constrained state machine.<n>On a recruitment system backed by the Beisen iTalent platform, 185 expert-curated scenarios trigger 1671 live API calls.<n>Our GSPO-aligned 7B Intent Router achieves higher joint accuracy than zero-shot GPT-4o on this FSM-constrained adversarial routing benchmark.
• Score: 0.0
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: Multi-agent orchestration frameworks such as LangChain, LangGraph, and CrewAI route tasks through graph-based pipelines but do not enforce the stage constraints that govern real business processes. We present SDOF, a framework that treats multi-agent execution as a constrained state machine. SDOF operates through two primary defensive layers, implemented by three components: (1) an Online-RLHF Specialized Intent Router trained via Generative Reward Modeling (GRPO) and (2) a StateAwareDispatcher with GoalStage finite-automaton checks and precondition/postcondition SkillRegistry validation for auditable execution control. On a recruitment system backed by the Beisen iTalent platform (6000+ enterprises), 185 expert-curated scenarios trigger 1671 live API calls. Our GSPO-aligned 7B Intent Router achieves higher joint accuracy than zero-shot GPT-4o on this FSM-constrained adversarial routing benchmark (80.9% versus 48.9%). In end-to-end execution, SDOF reaches 86.5% task completion (95% confidence interval 80.8 to 90.7) and blocks all 22 operations in the injection, illegal HR subset. Under a broader message-level blocking audit, SDOF attains precision 100% and recall 88%, expert agreement kappa=0.94. A separate evaluation on 960 SGD-derived dialogues spanning 8 service domains surfaces 201 stage-order conflicts under our FSM mapping, 41 of which arise in the normal split. This arXiv version reports the current validated scope; extended multi-seed training comparisons and deeper workflow evaluations will be released in a subsequent update.

Related papers

• From Production Traffic to Post-Training: Building a Self-Hosted LLM That Covers the Corporate Request Mix [55.08143827481755]
  We consolidate traffic from over 200 internal applications onto a single model.<n>We train a separate GRPO expert per axis and merge them via two-stage SLERP.<n>The model absorbs 50% of platform traffic, 116M requests per month, at a fraction of the serving cost.
  arXiv  Detail & Related papers  (2026-09-01T17:39:26Z)
• MasDrift: Benchmarking Authorization Preservation Across Multi-Agent Architectures [23.859862132370168]
  We introduce MasDrift, a benchmark of 600 benign productivity tasks across eight domains.<n>MasDrift compares single-agent, centralized, and decentralized coordination while varying hierarchy depth and peer width.<n>We compare two defenses that differ in where authorization evidence resides. One re-anchors every pending call to the original user request. The other carries an attenuated policy along the delegation chain.
  arXiv  Detail & Related papers  (2026-08-02T04:46:15Z)
• ClawArena-Team: Benchmarking Subagent Orchestration and Dynamic Workflows in Language-Model Agents [61.184954205350984]
  ClawArena-Team is a benchmark of 41 multi-turn, multimodal, multi-agent scenarios spanning 258 evaluation rounds and 72 staged updates.<n>All scoring is execution-based with no LLM judge: an overall score -- the Subagent-Management Score (SMS) -- multiplies task correctness by a least-privilege and modality-routing factor.
  arXiv  Detail & Related papers  (2026-06-30T06:06:08Z)
• Compliance-Scored Best-of-N Guardrail Orchestration for Multimodal Document Generation in Payments Dispute Defense [0.0]
  High-stakes enterprise document generation includes financial dispute narratives, compliance notices, and audit summaries.<n>Prior to a unified guardrail layer, production systems often stitched together separate PII redaction, content moderation, and format validation steps.<n>We present a guardrail orchestration layer for text and image inputs that couples multi-candidate generation with an explicit compliance score used for early exit.
  arXiv  Detail & Related papers  (2026-06-01T00:24:30Z)
• GTA: Generating Long-Horizon Tasks for Web Agents at Scale [82.43869456830664]
  We introduce a scalable framework, GTA, that integrates crawling, retrieval-based seeding, in-context generation, and automated quality control.<n>We instantiate the pipeline on over 50 websites covering e-commerce, government, forums, and news, with multilingual and multi-hop coverage.<n>Our contributions are three-fold: (i) formalizing multi-hop web-agent task generation, (ii) proposing an efficient and validated pipeline for automatic data creation, and (iii) releasing a dynamic benchmark with reproducible evaluation.
  arXiv  Detail & Related papers  (2026-05-28T01:05:50Z)
• CHI-Bench: Can AI Agents Automate End-to-End, Long-Horizon, Policy-Rich Healthcare Workflows? [150.84850629123287]
  End-to-end automation of realistic healthcare operations stresses three capabilities underrepresented in current benchmarks.<n>$$-Bench is a benchmark of long-horizon healthcare across three domains: provider prior authorization, payer utilization management, and care management.<n>Across 30 agent harness/models configurations, the best agent resolves only 28.0% of tasks, no agent clears 20% on strict pass3, and executing all tasks in a single session slumps the performance to 3.8%.
  arXiv  Detail & Related papers  (2026-05-15T22:34:31Z)
• ClawForge: Generating Executable Interactive Benchmarks for Command-Line Agents [59.626170560327274]
  We present textbfClawForge, a generator-backed benchmark framework for executable command-line categories under state conflict.<n>We instantiate this framework as the ClawForge-Bench (17 scenarios, 6 ability categories)
  arXiv  Detail & Related papers  (2026-05-13T21:34:08Z)
• Beyond Task Success: Measuring Workflow Fidelity in LLM-Based Agentic Payment Systems [1.8717456484053328]
  Agentic Success Rate (ASR) is a trajectory-fidelity metric that compares observed and expected agent execution sequences.<n>ASR is applied to the Hierarchical Multi-Agent System for Payments (HMASP) across 18 LLMs and 90,000 task instances.<n>Prompt refinements and deterministic routing guards guided by ASR diagnostics yield substantial TSR improvements.
  arXiv  Detail & Related papers  (2026-05-07T15:50:26Z)
• Measuring the Permission Gate: A Stress-Test Evaluation of Claude Code's Auto Mode [9.116800340266066]
  Claude Code's auto mode is the first deployed permission system for AI coding agents.<n>Anthropic reports a 0.4% false positive rate and 17% false negative rate on production traffic.<n>We evaluate 253 state-changing actions at the individual action level against oracle ground truth.
  arXiv  Detail & Related papers  (2026-04-04T17:56:30Z)
• AVDA: Autonomous Vibe Detection Authoring for Cybersecurity [0.1633272850273525]
  We introduce AVDA, a framework that leverages the Model Context Protocol (MCP) to automate detection by integrating organizational context into AI-assisted code generation.<n>We evaluate three authoring strategies - Baseline, Sequential, and Agentic - across a diverse corpus of production detections and state-of-the-art LLMs.<n>Our results show that Agentic attains 87% of Agentic quality at 40x lower token cost.
  arXiv  Detail & Related papers  (2026-03-26T21:52:33Z)
• SKILLS: Structured Knowledge Injection for LLM-Driven Telecommunications Operations [0.0]
  General-purpose language model (LLM) agents execute telecom operations through real API interfaces, or do they require structured domain guidance?<n>We introduce SKILLS, a benchmark framework comprising 37 telecom operations scenarios spanning 8 TM Forum Open API domains.<n>Results across 5 open-weight model conditions and 185 scenario-runs show consistent skill lift across all models.
  arXiv  Detail & Related papers  (2026-03-16T14:48:53Z)
• Towards a Science of Scaling Agent Systems [79.64446272302287]
  We formalize a definition for agent evaluation and characterize scaling laws as the interplay between agent quantity, coordination structure, modelic, and task properties.<n>We derive a predictive model using coordination metrics, that cross-validated R2=0, enabling prediction on unseen task domains.<n>We identify three effects: (1) a tool-coordination trade-off: under fixed computational budgets, tool-heavy tasks suffer disproportionately from multi-agent overhead, and (2) a capability saturation: coordination yields diminishing or negative returns once single-agent baselines exceed 45%.
  arXiv  Detail & Related papers  (2025-12-09T06:52:21Z)
• SOPBench: Evaluating Language Agents at Following Standard Operating Procedures and Constraints [59.645885492637845]
  SOPBench is an evaluation pipeline that transforms each service-specific SOP code program into a directed graph of executable functions.<n>Our approach transforms each service-specific SOP code program into a directed graph of executable functions and requires agents to call these functions based on natural language SOP descriptions.<n>We evaluate 18 leading models, and results show the task is challenging even for top-tier models.
  arXiv  Detail & Related papers  (2025-03-11T17:53:02Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.