{"paper":{"title":"Understanding Bugs in Modern Agentic Frameworks: A Study of Symptoms, Root Causes, and Triggering Conditions","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Analysis of 409 fixed bugs shows agentic frameworks have distinct failure modes from autonomous orchestration.","cross_cats":[],"primary_cat":"cs.SE","authors_text":"Hannuo Zhang, Shin Hwei Tan, Xiaowen Zhang","submitted_at":"2026-04-10T03:13:24Z","abstract_excerpt":"Modern agentic frameworks (e.g., CrewAI and AutoGen) have evolved into complex, autonomous multi-agent systems, introducing unique reliability challenges beyond earlier pipeline-based LLM libraries. However, existing empirical studies focus on earlier LLM libraries or task-level bugs, leaving the unique complexities of these agentic frameworks unexplored. We bridge the gap by conducting a comprehensive study of 409 fixed bugs from five representative agentic frameworks. We propose a five-layer abstraction to capture structural complexities in agentic frameworks, spanning from orchestration to "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Our study uncovers specialized symptoms, such as unexpected execution sequences and user configurations ignored, which are unique to autonomous orchestration. We further identify agent-specific root causes, including model-related faults, cognitive context mismanagement, and orchestration faults. Statistical analysis reveals cross-framework consistency and significant associations among these bug dimensions.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The assumption that the 409 fixed bugs from the five selected frameworks are representative of the full range of issues in modern agentic systems and that the proposed five-layer abstraction fully captures their structural complexities.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Analysis of bugs in modern agentic frameworks uncovers unique symptoms like unexpected execution sequences and root causes including model faults and orchestration issues, with transferable patterns across designs.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Analysis of 409 fixed bugs shows agentic frameworks have distinct failure modes from autonomous orchestration.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"d6681780764ffe570b74c85d48e6c3a3acf2e884978311c0d5d2998629a6382b"},"source":{"id":"2604.08906","kind":"arxiv","version":2},"verdict":{"id":"f353c799-4600-4631-b7c7-1bb090789086","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T17:44:29.153567Z","strongest_claim":"Our study uncovers specialized symptoms, such as unexpected execution sequences and user configurations ignored, which are unique to autonomous orchestration. We further identify agent-specific root causes, including model-related faults, cognitive context mismanagement, and orchestration faults. Statistical analysis reveals cross-framework consistency and significant associations among these bug dimensions.","one_line_summary":"Analysis of bugs in modern agentic frameworks uncovers unique symptoms like unexpected execution sequences and root causes including model faults and orchestration issues, with transferable patterns across designs.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The assumption that the 409 fixed bugs from the five selected frameworks are representative of the full range of issues in modern agentic systems and that the proposed five-layer abstraction fully captures their structural complexities.","pith_extraction_headline":"Analysis of 409 fixed bugs shows agentic frameworks have distinct failure modes from autonomous orchestration."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.08906/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}