{"paper":{"title":"SkillMOO: Multi-Objective Optimization of Agent Skills for Software Engineering","license":"http://creativecommons.org/licenses/by/4.0/","headline":"SkillMOO evolves skill bundles for LLM coding agents by combining LLM-proposed edits with NSGA-II selection to raise pass rates while lowering cost.","cross_cats":["cs.AI"],"primary_cat":"cs.SE","authors_text":"Alina Geiger, Dominik Sobania, Federica Sarro, Jie M. Zhang, Jingzhi Gong, Lukas Twist, Ruizhen Gu, Shuo Han, Yazhuo Cao, Zhiwei Fei","submitted_at":"2026-04-10T13:08:01Z","abstract_excerpt":"Agent skills are increasingly used to configure coding agents for software engineering (SE) tasks, yet current practice treats them as static, hand-crafted assets, or evolved on pass rate alone. This is insufficient: a skill can improve task success while substantially raising token cost, or introducing misleading guidance. We argue that SE agent skill bundles can be treated as multi-objective search objects and present SkillMOO, a framework that evolves skill bundles through LLM-proposed edits and NSGA-II Pareto selection on pass rate and inference cost. Evaluated across all 16 SkillsBench SE"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"On three SkillsBench software engineering tasks, SkillMOO improves pass rate by up to 131% while reducing cost up to 32% relative to the best baseline per task at low optimization overhead.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That LLM-proposed edits guided by failure analysis combined with NSGA-II survivor selection will reliably discover superior skill bundles without overfitting to the specific benchmark tasks or introducing hidden costs not captured in the reported metrics.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"SkillMOO automatically evolves skill bundles for LLM coding agents via LLM-proposed edits and NSGA-II, achieving up to 131% higher pass rates and 32% lower costs on three SkillsBench tasks.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"SkillMOO evolves skill bundles for LLM coding agents by combining LLM-proposed edits with NSGA-II selection to raise pass rates while lowering cost.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"cbcacd44beb112a8b4be55d46a21845dd6cf79d31e2ae3cfe8686f4c24a705b4"},"source":{"id":"2604.09297","kind":"arxiv","version":2},"verdict":{"id":"d8720abe-ad4d-44d5-9df5-d4da9c5e24a1","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T17:21:55.619579Z","strongest_claim":"On three SkillsBench software engineering tasks, SkillMOO improves pass rate by up to 131% while reducing cost up to 32% relative to the best baseline per task at low optimization overhead.","one_line_summary":"SkillMOO automatically evolves skill bundles for LLM coding agents via LLM-proposed edits and NSGA-II, achieving up to 131% higher pass rates and 32% lower costs on three SkillsBench tasks.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That LLM-proposed edits guided by failure analysis combined with NSGA-II survivor selection will reliably discover superior skill bundles without overfitting to the specific benchmark tasks or introducing hidden costs not captured in the reported metrics.","pith_extraction_headline":"SkillMOO evolves skill bundles for LLM coding agents by combining LLM-proposed edits with NSGA-II selection to raise pass rates while lowering cost."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.09297/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"}