{ "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": "2609.27378", "title": "Psychoacoustically Aligned Latent Smoothing for Adversarial Robustness of Full-Duplex Speech-to-Speech Dialogue Models", "summary": "PALS replaces an undefended residual-vector-quantized latent interface in full-duplex speech-to-speech models with anisotropic Gaussian noise injection shaped by local codebook covariance, trained via a KL consistency objective, as a defense against psychoacoustically masked adversarial perturbations. It adds no inference-time cost but requires training-time access to the model's latent codebook interface and its covariance statistics, plus retraining with the consistency loss. It can fail against adaptive attackers who optimize through the smoothing noise, offers only a limited certified radius well below observed empirical robustness, and its validation is confined to a Moshi-style architecture, leaving transferability to other speech stacks unproven.", "rationale": "The defense requires white-box training access to a full-duplex speech model's RVQ latent interface, which is unavailable to a reader constrained to third-party APIs or a single GPU without foundation-model training, though the threat model is worth tracking if they deploy open speech agents.", "technique": "Psychoacoustically aligned latent smoothing (PALS) for speech agents", "family": "seguranca_e_guardrails", "recommendation": "observar", "published": "2026-09-23", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2609.27378", "independent_implementations": 0, "total_implementations": 0, "stars": 0, "citations": 0, "claims": [ { "claim_id": "claim-01", "statement": "PALS replaces an undefended residual-vector-quantized latent interface in full-duplex speech-to-speech models with anisotropic Gaussian noise injection shaped by local codebook covariance, trained via a KL consistency objective, as a defense against psychoacoustically masked adversarial perturbations. It adds no inference-time cost but requires training-time access to the model's latent codebook interface and its covariance statistics, plus retraining with the consistency loss. It can fail against adaptive attackers who optimize through the smoothing noise, offers only a limited certified radius well below observed empirical robustness, and its validation is confined to a Moshi-style architecture, leaving transferability to other speech stacks unproven.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2609.27378", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "Against a Moshi-style full-duplex agent, PALS reduces adversarial targeted hijack success from up to 91.7% to 8.3%, response suppression (mute) to 11.2%, and jailbreak to 9.1%, with clean-quality degradation within 2.3%, plus a certified ellipsoidal latent robustness radius up to 0.616.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2609.27378", "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": "" }