跳到正文
TakoAPI
返回技能列表

evolution-state-analyzer

由 wanng-ide 提供

Coding Agents & IDEs
455下载量由 clawskills.sh 提供
0浏览量
0 个赞登录后即可为这个技能点赞
—暂无评分
你的评分

简介

Analyzes the evolution memory graph for stagnation patterns, recurring failures, and success plateaus.

描述

该技能的功能

Reads a memory_graph.jsonl file to analyze evolution cycles for stagnation patterns, failure clusters, and gene performance trends. Produces structured recommendations for adjusting future evolution runs.

Automates meta-analysis that would otherwise require manually parsing raw JSONL logs and computing aggregate statistics by hand.

何时使用

  • Detecting when an evolution loop has plateaued across many cycles
  • Identifying which genes consistently underperform below a success threshold
  • Grouping repeated failure reasons to surface systemic bugs
  • Reviewing score trends after a long batch of evolution runs
  • Getting prioritized next-step recommendations before starting a new cycle

示例工作流

输入

User asks: analyze the current evolution state and recommend what to do next

智能体
  1. 1

    Loads memory_graph.jsonl from the evolution working directory

  2. 2

    Calculates overall success rate and detects stagnation streaks by cycle count

  3. 3

    Ranks all genes by success rate and flags those below 0.5

  4. 4

    Clusters failure reasons across cycles to identify recurring systemic issues

  5. 5

    Returns a prioritized recommendation list with specific actions

输出

JSON report with total cycle count, overall success rate, stagnation flag, top genes ranked by efficacy, and concrete recommendations such as switching intent or deprecating low-performing genes

安装

clawhub install wanng-ide/evolution-state-analyzer

运行 CLI 命令,或把这段提示词粘贴到 Claude Code、Cursor、Codex 或你在用的任意编程智能体里。

链接