{ "schema_version": 1, "builder_plan": { "scenario": "A workflow needs specialist review, but extra agents may add more cost than useful work.", "change": "Test the proposed division of work against a single-agent implementation.", "baseline": "One agent with the same tools and an equal total token and time budget.", "steps": [ "Locate the paper's implementation and check its license, model support, data needs, and hardware requirements. Stop here if you cannot obtain the required artifacts.", "Save a small set of representative product tasks and expected outcomes. Keep evaluation tasks out of any training or tuning, and choose quality and cost limits before running.", "Run the baseline and the proposed change on identical inputs. Keep other settings fixed and repeat paired runs to expose variation.", "Save per-task outputs, failures, timings, and all usage costs. Review regressions before expanding the test." ], "measure": "Accepted outputs, coordination failures, wall-clock time, and total cost across all agents.", "decision": "Keep the change only if it meets your preselected quality and latency limits and improves the metric you care about after setup and operating costs. Otherwise retain the baseline. This is an exploration plan, not a validated recommendation." }, "builder_review": null, "arxiv_id": "2610.22280", "title": "LEASH: Uncertainty-Aware Weak-to-Strong Escalation and Oversight for Computer-Use Agents", "summary": "The method replaces always-on frontier-model execution (or random/heuristic escalation) with a router that runs a cheap weak agent by default and escalates to a strong model only when the weak model's hidden states resemble past failed trajectories, plus a shadow observer that monitors the black-box strong agent after handoff. It costs maintaining a bank of successful and failed trajectory activations, per-step hidden-state comparisons, and the engineering complexity of a three-layer routing pipeline, though the estimator reuses existing activations so no extra forward pass or monitor training is needed. It can fail if the weak model's latent uncertainty is miscalibrated for a new task distribution, if logged trajectories do not cover the deployment domain, or if the strong agent's errors are not detectable from black-box outputs alone.", "rationale": "Cost-aware escalation directly fits a constrained budget, but the reported gains are specific to OSWorld computer-use tasks and must be validated on the reader's own workload before adoption.", "technique": "Uncertainty-driven weak-to-strong model escalation", "family": "orquestracao_multiagente", "recommendation": "testar", "published": "2026-10-15", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2610.22280", "independent_implementations": 1, "total_implementations": 2, "stars": 132, "citations": null, "claims": [ { "claim_id": "claim-01", "statement": "The method replaces always-on frontier-model execution (or random/heuristic escalation) with a router that runs a cheap weak agent by default and escalates to a strong model only when the weak model's hidden states resemble past failed trajectories, plus a shadow observer that monitors the black-box strong agent after handoff. It costs maintaining a bank of successful and failed trajectory activations, per-step hidden-state comparisons, and the engineering complexity of a three-layer routing pipeline, though the estimator reuses existing activations so no extra forward pass or monitor training is needed. It can fail if the weak model's latent uncertainty is miscalibrated for a new task distribution, if logged trajectories do not cover the deployment domain, or if the strong agent's errors are not detectable from black-box outputs alone.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.22280", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "On OSWorld, LEASH with Qwen3.5-35B-A3B reaches 73.66% task success at $0.4117 per task, matching always-on Claude Opus 5 (72.90%) while reducing inference cost by 19.2%; it outperforms random escalation by 6.55-11.94 success points at matched escalation rates.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.22280", "source_page": null, "source_excerpt": "" } ], "exposure_map": [ { "dimension": "quality", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "compute", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "latency", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "operations", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "compatibility", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "security", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "data_and_training", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "reproducibility", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" } ], "risks": [], "minimum_test": [], "open_questions": [], "independent_tests": [], "lab_reproductions": [], "report_available": false, "validation_tier": "unknown", "evidence_tier": "unknown", "infrastructure_basis": "unknown", "equations": [], "equations_status": "not_fetched", "core_kind": "", "equations_fetched_at": "", "equations_source_url": "", "equations_source_sha256": "", "exposures_reviewed_at": "", "exposures_source_sha256": "" }