PulseAugur
实时 14:21:00
English(EN) Enabling Bash costs more context than seven MCP tool schemas

AI 代理中的 Bash 工具集成比预期产生更高的令牌成本

最近一项比较不同 AI 代理交互方法的分析显示,启用 Bash 命令可能比预期更消耗令牌。研究发现,在代理提示中使用 Bash 工具大约需要 2,700 个令牌,这显著高于七个 MCP 工具模式的总成本。这种更高的成本归因于 Bash 所需的绑定提示的开销。研究还强调,每个回合急切地加载多个 MCP 服务器是最昂贵的因素,其成本超过了 Bash 或工具模式本身的成本。 AI

影响 了解不同代理交互方法的令牌成本可以为 AI 应用程序的优化策略提供信息。

排序理由 对 AI 代理交互成本和令牌效率的分析。[lever_c_demoted from research: ic=1 ai=1.0]

在 dev.to — MCP tag 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →

AI 代理中的 Bash 工具集成比预期产生更高的令牌成本

本文如何被排名

Signal score
54 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
对 AI 代理交互成本和令牌效率的分析。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
infra, product
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准

报道来源 [1]

  1. dev.to — MCP tag TIER_1 English(EN) · Bryan Clark ·

    启用 Bash 比七个 MCP 工具模式需要更多上下文

    <blockquote> <p>Nine days ago I measured MCP tool schemas against a <code>curl</code> recipe with a tokenizer and got a near-tie. This time I ran the agent: four arms, same model, same 12 asks, only the delivery mechanism varying. <strong>MCP was the cheapest arm</strong> — 109,7…