[2605.04107] TSCG: Deterministic Tool-Schema Compilation for Agentic LLM Deployments

Computer Science > Software Engineering arXiv:2605.04107 (cs) [Submitted on 4 May 2026] Title:TSCG: Deterministic Tool-Schema Compilation for Agentic LLM Deployments Authors:Furkan Sakizli View a PDF of the paper titled TSCG: Deterministic Tool-Schema Compilation for Agentic LLM Deployments, by Furkan Sakizli View PDF HTML (experimental) Abstract:Production agent frameworks (OpenAI Function Calling, Anthropic Tool Use, MCP) transmit tool schemas as JSON, a format designed for machine parsing, not for interpretation by language models. For small models (4B-14B), this protocol mismatch accounts for the majority of tool-use failure at production catalog sizes. We present TSCG, a deterministic tool-schema compiler that resolves this mismatch at the API boundary, converting JSON schemas into token-efficient structured text without model access, fine-tuning, or runtime search. TSCG combines eight composable operators with a formal compression bound (>=51% on well-formed schemas). On TSCG-Agentic-Bench (about 19,000 calls, 12 models, 5 scenarios), TSCG restores Phi-4 14B from 0% to 84.4% accuracy at 20 tools (90.3% at 50 tools) and achieves 108-181% accuracy-retained ratio across three models on BFCL. Format-versus-compression decomposition (R^2=0.88 -> 0.03) establishes representation change as the dominant mechanism. Per-operator isolation across three frontier models reveals three distinct operator-response profiles: operator-hungry (Opus 4.7), operator-sensitive (GPT-5.2), an...