{ "schema_version": 1, "builder_plan": { "scenario": "An agent reads external documents or tool responses that may contain malicious instructions.", "change": "Replay attacks and legitimate lookalikes in an isolated copy of your agent workflow.", "baseline": "Your existing checks with the same tasks, permissions, and tool responses.", "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": "Successful attacks missed, legitimate actions blocked, detection delay, and review cost.", "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": "2610.00850", "title": "AuraForge: Scaling Security Supervision for Training Coding Agents", "summary": "It replaces reliance on scarce human-written security tests in real repositories with automatically synthesized and validated executable attack-oriented tests used as training supervision for coding agents. The cost is a substantial task-construction and validation pipeline (repository execution environments, anti-reward-hacking safeguards) plus the fine-tuning run itself, which is feasible for a 4B model on a 24 GB GPU but the gym-building infrastructure is not turnkey. Synthesized tests can still mislabel secure alternative implementations or overfit to the attack patterns the generator knows, leaving uncovered CWEs without supervision.", "rationale": "The reader cannot reproduce the data-synthesis pipeline cheaply, but can apply the validated principle—training or evaluating coding agents against executable security tests rather than functional correctness alone—at small scale and should validate it on their own workload.", "technique": "Synthesized executable security tests for agent training", "family": "seguranca_e_guardrails", "recommendation": "testar", "published": "2026-10-01", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2610.00850", "independent_implementations": 1, "total_implementations": 2, "stars": 14, "citations": 0, "claims": [ { "claim_id": "claim-01", "statement": "It replaces reliance on scarce human-written security tests in real repositories with automatically synthesized and validated executable attack-oriented tests used as training supervision for coding agents. The cost is a substantial task-construction and validation pipeline (repository execution environments, anti-reward-hacking safeguards) plus the fine-tuning run itself, which is feasible for a 4B model on a 24 GB GPU but the gym-building infrastructure is not turnkey. Synthesized tests can still mislabel secure alternative implementations or overfit to the attack patterns the generator knows, leaving uncovered CWEs without supervision.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.00850", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "AuraForge yields about 3x as many security test cases as human-written tests, reduces false-positive rate by 83.23%, and training Qwen3.5-4B on them improves SecPass by 6.2 points versus 4.4 with human tests.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.00850", "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": "" }