{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,9]],"date-time":"2026-07-09T06:13:04Z","timestamp":1783577584869,"version":"3.55.0"},"reference-count":23,"publisher":"IEEE","license":[{"start":{"date-parts":[[2026,6,22]],"date-time":"2026-06-22T00:00:00Z","timestamp":1782086400000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,6,22]],"date-time":"2026-06-22T00:00:00Z","timestamp":1782086400000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,6,22]]},"DOI":"10.1109\/dsn-w70714.2026.00022","type":"proceedings-article","created":{"date-parts":[[2026,7,8]],"date-time":"2026-07-08T19:42:00Z","timestamp":1783539720000},"page":"9-16","source":"Crossref","is-referenced-by-count":0,"title":["From Agent Failure Paths to Quantified Residual Risk: A Compositional Framework for Resilient Agentic AI"],"prefix":"10.1109","author":[{"given":"Hassan","family":"Karim","sequence":"first","affiliation":[{"name":"Stable Cyber,Dallas,TX,USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sai","family":"Sitharaman","sequence":"additional","affiliation":[{"name":"Zetafence,Dublin,CA,USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Deepti","family":"Gupta","sequence":"additional","affiliation":[{"name":"Texas A&M University–Central Texas,Killeen,TX,USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Danda B.","family":"Rawat","sequence":"additional","affiliation":[{"name":"Howard University,Washington, DC,USA"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref1","article-title":"Gartner predicts over 40% of agentic AI projects will be canceled by end of 2027","year":"2025"},{"key":"ref2","article-title":"Sok: The attack surface of agentic ai \u2013 tools, and autonomy","author":"Dehghantanha","year":"2026"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1016\/S0925-7535(03)00047-X"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1016\/j.jisa.2016.05.008"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.21314\/jor.2000.038"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1111\/risa.13331"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1093\/logcom\/exs029"},{"key":"ref8","article-title":"DSPy: Compiling declarative language model calls into self-improving pipelines","volume-title":"Proc. Int. Conf. Learning Representations (ICLR)","author":"Khattab"},{"key":"ref9","article-title":"AutoGen: Enabling next-gen LLM applications via multi-agent conversation","volume-title":"Proc. Conf. Language Modeling (COLM)","author":"Wu"},{"key":"ref10","article-title":"MetaGPT: Meta programming for a multi-agent collaborative framework","volume-title":"Proc. Int. Conf. Learning Representations (ICLR)","author":"Hong"},{"key":"ref11","article-title":"Why do multi-agent LLM systems fail?","author":"Cemri","year":"2025","journal-title":"Advances in Neural Information Processing Systems (NeurIPS), Datasets and Benchmarks Track"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1111\/1467-9965.00068"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1016\/S0266-8920(01)00019-4"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1038\/nature08932"},{"key":"ref15","article-title":"Agentspec: Customizable runtime enforcement for safe and reliable llm agents","author":"Wang","year":"2025"},{"key":"ref16","article-title":"Pro2guard: Proactive runtime enforcement of LLM agent safety via probabilistic model checking","author":"Wang","year":"2026"},{"key":"ref17","article-title":"Defeating prompt injections by design","author":"Debenedetti","year":"2025"},{"key":"ref18","doi-asserted-by":"crossref","DOI":"10.14722\/ndss.2025.241131","article-title":"Isolategpt: An execution isolation architecture for llm-based agentic systems","author":"Wu","year":"2025"},{"key":"ref19","article-title":"Progent: Programmable privilege control for llm agents","author":"Shi","year":"2025"},{"key":"ref20","article-title":"Assurance of AI systems from a dependability perspective","author":"Bloomfield","year":"2024","journal-title":"SRI International, Tech. Rep. SRI-CSL-2024-02"},{"key":"ref21","first-page":"26 202","article-title":"On the resilience of LLM-based multi-agent collaboration with faulty agents","volume-title":"Proc. 42nd Int. Conf. Machine Learning (ICML), ser. PMLR","volume":"267","author":"Huang"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2025.3561235"},{"key":"ref23","article-title":"Forge-bench: A threat-labeled dataset and digital twin framework for security evaluation of LLM-driven warehouse robots","author":"Karim","year":"2026"}],"event":{"name":"2026 56th Annual IEEE International Conference on Dependable Systems and Networks Workshops (DSN-W)","location":"Charlotte, NC, USA","start":{"date-parts":[[2026,6,22]]},"end":{"date-parts":[[2026,6,25]]}},"container-title":["2026 56th Annual IEEE International Conference on Dependable Systems and Networks Workshops (DSN-W)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/11594430\/11594448\/11594618.pdf?arnumber=11594618","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,9]],"date-time":"2026-07-09T05:12:56Z","timestamp":1783573976000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11594618\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,6,22]]},"references-count":23,"URL":"https:\/\/doi.org\/10.1109\/dsn-w70714.2026.00022","relation":{},"subject":[],"published":{"date-parts":[[2026,6,22]]}}}