{"record": {"aliases": ["AgenticSTS: A Bounded-Memory Testbed for Long-Horizon LLM Agents"], "artifacts": [{"id": "arxiv:2607.02255", "kind": "paper", "url": "https://arxiv.org/abs/2607.02255"}], "authors": ["Xiangchen Cheng", "Yunwei Jiang", "Jianwen Sun", "Zizhen Li", "Chuanhao Li", "Xiangcheng Cao", "Yihao Liu", "Fanrui Zhang", "Li Jin", "Kaipeng Zhang"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2607.02255", "source_url": "https://arxiv.org/abs/2607.02255"}, "categories": ["Agents", "General AI", "Language & Knowledge"], "description": {"en": "AgenticSTS is a reproducible long-horizon agent testbed that evaluates bounded, typed-retrieval memory configurations through game outcomes and controlled ablations."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2607.02255", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2607.02255", "title": "AgenticSTS"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["Shanghai Jiao Tong University", "Avaya (Bermuda)", "Shanghai Innovation Institute", "University of Science and Technology of China"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2607.02255", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2607.02255", "source_url": "https://openalex.org/W7167282465"}, "key": "claire-radar:2607.02255", "modality": null, "name": "AgenticSTS", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"admission_policy_version": "2026-09-09.1", "confidence": "Medium", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-09-25T04:40:22Z", "data_status": "primary-source-indexed", "display_eligible": "true", "origin_record_id": "bm_agenticsts-a-bounded-memory-testbed-for-lo_d672ecae", "origin_source": "arxiv", "recognition_confidence": "0.6", "record_sha256": "3ba694e88228ca3f652ff0f5f6203f232b1dac91b8edc35540c28608a5eb1e60", "relation": "introduces", "review_model": "codex-cli-default", "review_state": "ai-reviewed", "reviewed_at": "2026-09-24T19:01:37.943600Z", "source_url": "https://arxiv.org/abs/2607.02255"}, "publisher": null, "released": "2026-07-02", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2607.02255", "source_url": "https://arxiv.org/abs/2607.02255"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2607-02255", "source": "claire_radar", "source_benchmark_id": "2607.02255", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": ["AgenticSTS: A Bounded-Memory Testbed for Long-Horizon LLM Agents"], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Language & Knowledge", "attentionForecast": {"confidence": "Medium", "horizon": "7d", "reason": "Long-horizon memory evaluation and released trajectories are likely to interest agent researchers, although the reported controlled sample is small.", "score": 68}, "capabilities": [], "confidence": "Medium", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-09-24T18:59:59.467212Z", "inputHash": "9889319284bd773ab3c66d437d4ea8f9d3c653b48416c119c8e54433867555c0", "model": "codex-cli-default"}, "curation": {"admissionPolicyVersion": "2026-09-09.1", "canonicalNameEvidence": "AgenticSTS: A Bounded-Memory Testbed for Long-Horizon LLM Agents", "canonicalNameSource": "paper_title", "decisionReason": "The supplied text names the testbed, defines the long-horizon task and win-based comparison, and states that trajectories, frozen snapshots, prompt records, and analysis scripts are released.", "evaluationEvidence": {"reuse": "We release a reproducible testbed: 298 completed trajectories with condition tags, frozen memory/skill snapshots, prompt records, and analysis scripts", "scoring": "the no-store baseline wins 3/10 games and adding the skill layer 6/10", "task": "We instantiate the contract in Slay the Spire 2, a closed-rule stochastic deck-building game whose runs require hundreds of tactical and strategic decisions."}, "model": "codex-cli-default", "reviewedAt": "2026-09-24T19:01:37.943600Z", "state": "ai-reviewed"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "AgenticSTS is a reproducible long-horizon agent testbed that evaluates bounded, typed-retrieval memory configurations through game outcomes and controlled ablations.", "displayEligible": true, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "public_reusable", "evidence": {"reasonCodes": ["benchmark term in abstract", "benchmark release stated in one sentence", "evaluation protocol evidence"], "snippet": "We release a reproducible testbed: 298 completed trajectories with condition tags, frozen memory/skill snapshots, prompt records, and analysis scripts -- an agent design and a validated, reusable methodology for studying how explicit memory layers shape long-horizon LLM-agent decisions."}, "familyId": "bmf_7649d75673dd", "firstSeenAt": "2026-08-19", "heat": null, "id": "bm_agenticsts-a-bounded-memory-testbed-for-lo_d672ecae", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": null, "data": null, "paper": "https://arxiv.org/abs/2607.02255", "pdf": "https://arxiv.org/pdf/2607.02255", "project": null, "report": "https://arxiv.org/abs/2607.02255"}, "metrics": [], "motivation": "Memory for a long-horizon LLM agent is a contract about what each future decision is allowed to see.", "name": "AgenticSTS", "oneLine": "AgenticSTS is a reproducible long-horizon agent testbed that evaluates bounded, typed-retrieval memory configurations through game outcomes and controlled ablations.", "paperTitle": "AgenticSTS: A Bounded-Memory Testbed for Long-Horizon LLM Agents", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2607.02255", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-08-19T11:05:13.395391Z"}, "publications": [], "readiness": "Paper only", "recognitionConfidence": 0.6, "relation": "introduces", "releasedAt": "2026-07-02", "reviewLifecycle": {"state": "pending"}, "source": {"authors": ["Xiangchen Cheng", "Yunwei Jiang", "Jianwen Sun", "Zizhen Li", "Chuanhao Li", "Xiangcheng Cao", "Yihao Liu", "Fanrui Zhang", "Li Jin", "Kaipeng Zhang"], "categories": ["cs.AI", "cs.CL"], "id": "2607.02255", "title": "AgenticSTS: A Bounded-Memory Testbed for Long-Horizon LLM Agents", "type": "arxiv", "url": "https://arxiv.org/abs/2607.02255"}, "sourceUpdatedAt": "2026-07-02T14:44:32Z", "topics": ["Agents"], "venueAttempts": [], "whyItMatters": "It supports repeatable investigation of how explicit memory layers affect decisions across hundreds of steps."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
