{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:P7JA5REZLSS5UVXEB4SDK6GF22","short_pith_number":"pith:P7JA5REZ","schema_version":"1.0","canonical_sha256":"7fd20ec4995ca5da56e40f243578c5d6b3b52acfcef72ce5f8c8830987e7fe02","source":{"kind":"arxiv","id":"2606.03005","version":1},"attestation_state":"computed","paper":{"title":"MUSE: A Unified Agentic Harness for MLLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Hailing Wang, Jianglin Lu, Mingyuan Zhang, Qihua Dong, Xu Ma, Yizhou Wang, Yun Fu","submitted_at":"2026-06-02T01:24:30Z","abstract_excerpt":"Despite rapid progress, multimodal large language models (MLLMs) still fail on tasks that humans solve effortlessly, such as navigating a grid maze from a screenshot or selecting the correct puzzle piece. Rather than retraining the model, we ask a complementary question: how much capability can be elicited from a frozen MLLM purely by improving the execution scaffold around it? We introduce MUSE, a multimodal unified structured execution harness that wraps any off-the-shelf MLLM with composable modules for task representation, visual processing, perception tool use, structured parsing, determi"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2606.03005","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-06-02T01:24:30Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"e14f9aad6040f8845ffe4d3ff6f797f6bb6b1ce0c939f220a7c06f108bd6dfd0","abstract_canon_sha256":"9aff33f986572f441d16010fca5841ced4e8b1b609b6880a045deb4420d5087e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-03T01:05:29.068595Z","signature_b64":"qbO3+ypcWVRRnCCrjtrj0Zk+ITseT3NEgLp312sCgye3qdJhIm/owTl+35zXeCEJlpa+vZYTzYPkYciz2re+BA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7fd20ec4995ca5da56e40f243578c5d6b3b52acfcef72ce5f8c8830987e7fe02","last_reissued_at":"2026-06-03T01:05:29.068153Z","signature_status":"signed_v1","first_computed_at":"2026-06-03T01:05:29.068153Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MUSE: A Unified Agentic Harness for MLLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Hailing Wang, Jianglin Lu, Mingyuan Zhang, Qihua Dong, Xu Ma, Yizhou Wang, Yun Fu","submitted_at":"2026-06-02T01:24:30Z","abstract_excerpt":"Despite rapid progress, multimodal large language models (MLLMs) still fail on tasks that humans solve effortlessly, such as navigating a grid maze from a screenshot or selecting the correct puzzle piece. Rather than retraining the model, we ask a complementary question: how much capability can be elicited from a frozen MLLM purely by improving the execution scaffold around it? We introduce MUSE, a multimodal unified structured execution harness that wraps any off-the-shelf MLLM with composable modules for task representation, visual processing, perception tool use, structured parsing, determi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.03005","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2606.03005/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2606.03005","created_at":"2026-06-03T01:05:29.068222+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.03005v1","created_at":"2026-06-03T01:05:29.068222+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.03005","created_at":"2026-06-03T01:05:29.068222+00:00"},{"alias_kind":"pith_short_12","alias_value":"P7JA5REZLSS5","created_at":"2026-06-03T01:05:29.068222+00:00"},{"alias_kind":"pith_short_16","alias_value":"P7JA5REZLSS5UVXE","created_at":"2026-06-03T01:05:29.068222+00:00"},{"alias_kind":"pith_short_8","alias_value":"P7JA5REZ","created_at":"2026-06-03T01:05:29.068222+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/P7JA5REZLSS5UVXEB4SDK6GF22","json":"https://pith.science/pith/P7JA5REZLSS5UVXEB4SDK6GF22.json","graph_json":"https://pith.science/api/pith-number/P7JA5REZLSS5UVXEB4SDK6GF22/graph.json","events_json":"https://pith.science/api/pith-number/P7JA5REZLSS5UVXEB4SDK6GF22/events.json","paper":"https://pith.science/paper/P7JA5REZ"},"agent_actions":{"view_html":"https://pith.science/pith/P7JA5REZLSS5UVXEB4SDK6GF22","download_json":"https://pith.science/pith/P7JA5REZLSS5UVXEB4SDK6GF22.json","view_paper":"https://pith.science/paper/P7JA5REZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.03005&json=true","fetch_graph":"https://pith.science/api/pith-number/P7JA5REZLSS5UVXEB4SDK6GF22/graph.json","fetch_events":"https://pith.science/api/pith-number/P7JA5REZLSS5UVXEB4SDK6GF22/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/P7JA5REZLSS5UVXEB4SDK6GF22/action/timestamp_anchor","attest_storage":"https://pith.science/pith/P7JA5REZLSS5UVXEB4SDK6GF22/action/storage_attestation","attest_author":"https://pith.science/pith/P7JA5REZLSS5UVXEB4SDK6GF22/action/author_attestation","sign_citation":"https://pith.science/pith/P7JA5REZLSS5UVXEB4SDK6GF22/action/citation_signature","submit_replication":"https://pith.science/pith/P7JA5REZLSS5UVXEB4SDK6GF22/action/replication_record"}},"created_at":"2026-06-03T01:05:29.068222+00:00","updated_at":"2026-06-03T01:05:29.068222+00:00"}