{ "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": "2609.19425", "title": "Closed-World Resolution Against Tool Hallucination in LLM Agents", "summary": "The resolver replaces nothing in the existing pipeline; it adds a training-free pre-gate check (tool registry membership plus argument signature validation) that rejects calls to nonexistent tools or with undeclared arguments before any causal or security gate runs, since a hallucinated call is by construction outside what a gate decided. Its cost is negligible compute and latency (a lookup and a schema check), with the engineering burden of maintaining an accurate, versioned tool registry and correctly handling merged MCP namespaces where name collisions and shadowing create surfaces a single registry cannot express. It fails on the one irreducible residue: calls to real tools with borrowed arguments that are schema-indistinguishable from valid calls, which requires downstream gating or execution checks rather than closed-world resolution.", "rationale": "It is a cheap, training-free validation layer that fits a 24 GB GPU or API-only setup and measurably catches a failure class that tool selection and gating structurally cannot.", "technique": "Closed-world tool-call resolver before gating", "family": "seguranca_e_guardrails", "recommendation": "adotar", "published": "2026-09-16", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2609.19425", "independent_implementations": 1, "total_implementations": 2, "stars": 0, "citations": null, "claims": [ { "claim_id": "claim-01", "statement": "The resolver replaces nothing in the existing pipeline; it adds a training-free pre-gate check (tool registry membership plus argument signature validation) that rejects calls to nonexistent tools or with undeclared arguments before any causal or security gate runs, since a hallucinated call is by construction outside what a gate decided. Its cost is negligible compute and latency (a lookup and a schema check), with the engineering burden of maintaining an accurate, versioned tool registry and correctly handling merged MCP namespaces where name collisions and shadowing create surfaces a single registry cannot express. It fails on the one irreducible residue: calls to real tools with borrowed arguments that are schema-indistinguishable from valid calls, which requires downstream gating or execution checks rather than closed-world resolution.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2609.19425", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "Across ten hosted models, fabricated-tool calls concentrate on the unconstrained raw-JSON invocation surface (34 hallucinations vs. 3 on the constrained surface); model scale does not reduce hallucinations (a 675B model matches a 7-8B model).", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2609.19425", "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": "" }