Evidence record EV-0010

Agent scaffolding, not MCP versus CLI, drove cost

A controlled comparison (preprint) found that on one git task the agent scaffolding drove cost more than MCP versus CLI: CLI-only scaffoldings were 5.0x to 28x cheaper, and paired MCP-to-CLI cost ratios ranged from 0.43x to 29x.

Preprint · Measurement · Retrieved 8 October 2026

Evidence class: Preprint. Unreviewed. Many are by one author or a small team, and some authors have a stake in the result. Most AX research is in this class today. Pattern tags: cost, evaluation, measurement.

Effect, as the source reports it

Cost per run, scaffoldings without MCP support versus those with it (CLI runs only). Baseline: Five scaffoldings that support MCP. With the change: Two scaffoldings with no MCP support. Direction: decrease. Size: 5.0x to 28x cheaper. Sample: Seven scaffoldings; five models; one fixed task of six git operations.
Paired MCP-to-CLI cost ratio. Direction: mixed. Size: Thirteen strictly paired ratios span 0.43x to 29x. Sample: 13 pairs.
Share of spend that bought no completed work. Baseline: CLI runs: 2.2%. With the change: MCP runs: 12.9%. Direction: increase. Size: 2.2% to 12.9%; failure frequency was the same for both. Sample: Original runs and repetitions.
Cost variation of a local 27-billion-parameter model across scaffoldings. Direction: not-applicable. Size: 139x. Sample: Seven scaffoldings.

Agent profile

Note on models: Five language models, including a local 27-billion-parameter model. The abstract does not name them.
Harness: Seven agent scaffoldings (not named in the abstract).
Extensions: MCP server for the git service (in MCP conditions).

Conflicts of interest

None declared in the abstract. The full text was not checked for a competing-interest statement.

Source

The Scaffolding Matters More Than the Interface: A Controlled Comparison of MCP and CLI Tool Use Across Seven Agent Scaffoldings, Five Language Models, and One Software Task, arXiv, Marc Alier Forment, María José Casañ Guerrero, Francisco José García-Peñalvo, Juanan Pereira, 9 August 2026, arXiv:2608.08654. Retrieved 8 October 2026; verification: abstract-only.

Every number in this record was checked against the live arXiv abstract page on 2026-10-08. The full text was not re-checked.

Limitations

Preprint, not peer reviewed.
One task against one kind of service.
Agents often ignored the interface they were assigned, so some comparisons mix interfaces.

For designers

Do not claim that MCP or a CLI is cheaper in general; the client running the agent mattered more here. When you test an interface, check what the agent actually called, because agents often ignored the one they were given.

ACT-07: There is a direct route, not only clicks

Cite this record

Cite the original source for any number, and keep the evidence class and model set with the figure. To point at this record, use "AX evidence register, EV-0010" and this page's address, https://agentexperience.tech/evidence/ev-0010/. The record is also in /evidence.json. The register's licence will be confirmed before its source repository is published.