Data and scripts for "Efficient Tool Use for Agentic AI: A Survey and an Analytical Cost Model for Tool-Using LLM Agents"

Supporting data and scripts for the survey Efficient Tool Use for Agentic AI (H. Mohasel Arjomandi, preprint, November 2026; survey cutoff October 31, 2026): arXiv harvest and screening sheet, coverage-check decisions, full-text audit of 43 tool-use benchmarks, outputs of the cost-model consistency check on public tau-bench and SWE-agent traces and of the BFCL catalog-padding sweep, and the scripts that recompute every number of the cost model, bibliometrics and audit. Third-party traces are not redistributed; download locations are in the README. Scripts: MIT. Data: CC BY 4.0.