AgentSkillsCN

token-economist

通过优化数据输入，最大限度降低令牌消耗与成本；对大型文件进行智能摘要与分块处理，同时确保关键上下文信息不被遗漏。

SKILL.md
--- frontmatter
name: token-economist
description: Minimizes token consumption and costs by optimizing data input. Performs smart summarization and chunking of large files without losing critical context.
status: implemented

Token Economist

This skill focuses on cost-efficiency and precision when dealing with large-scale data in LLMs.

Capabilities

1. Smart Summarization

  • •Condenses large log files or documentation into "information-dense" summaries.
  • •Strips redundant or boilerplate text before sending to the model.

2. Intelligent Chunking

  • •Splits large files based on logical boundaries (e.g., function blocks, sections) rather than raw character counts.
  • •Maintains cross-chunk context through small overlaps or "context headers."

Usage

  • •"Summarize this 10MB log file for analysis, keeping only the error patterns and stack traces."
  • •"The codebase is too large; use token-economist to prepare a dense overview for the AI."

Knowledge Protocol

  • •This skill adheres to the knowledge/orchestration/knowledge-protocol.md. It automatically integrates Public, Confidential (Company/Client), and Personal knowledge tiers, prioritizing the most specific secrets while ensuring no leaks to public outputs.