{ "schema_version": 1, "builder_plan": { "scenario": "Your coding assistant makes changes that still need substantial repair before merging.", "change": "Try the technique on held-out issues from a repository with reliable tests.", "baseline": "The same coding agent, repository snapshot, tools, and token 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": "Correct patches, regression failures, human repair time, and total cost per accepted change.", "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": "2609.39957", "title": "Learning When and How to Intervene: A Hindsight-Distilled Sentinel for Coding Agents", "summary": "Replaces reactive execution-feedback recovery and ad-hoc rule-based action checks with a lightweight 0.6B/1.7B classifier that, before each agent action, decides to allow, redirect, or escalate to a human and supplies corrective feedback. Costs an extra small-model inference per action plus the effort of obtaining or reproducing the SWE-Intervene hindsight-labeled trajectories and training the student, adding latency and an orchestration dependency to the agent loop. Can fail if the sentinel's intervention judgments do not transfer to the reader's repositories, task distribution, or agent family, and false interventions may block correct actions or inflate human escalation load.", "rationale": "The sentinel models are small enough for a single 24 GB GPU and the reported quality gains are meaningful, but adoption requires validating that hindsight-distilled intervention labels transfer to the reader's specific coding-agent workload.", "technique": "Hindsight-distilled pre-execution intervention sentinel", "family": "agentes_de_codigo", "recommendation": "testar", "published": "2026-09-30", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2609.39957", "independent_implementations": 0, "total_implementations": 0, "stars": 0, "citations": null, "claims": [ { "claim_id": "claim-01", "statement": "Replaces reactive execution-feedback recovery and ad-hoc rule-based action checks with a lightweight 0.6B/1.7B classifier that, before each agent action, decides to allow, redirect, or escalate to a human and supplies corrective feedback. Costs an extra small-model inference per action plus the effort of obtaining or reproducing the SWE-Intervene hindsight-labeled trajectories and training the student, adding latency and an orchestration dependency to the agent loop. Can fail if the sentinel's intervention judgments do not transfer to the reader's repositories, task distribution, or agent family, and false interventions may block correct actions or inflate human escalation load.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2609.39957", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "Improves task completion by up to 14% on SWE-bench Verified Mini and 10% on Ask or Assume across sentinel scales and coding-agent families, with competitive token consumption.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2609.39957", "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": "" }