菜单
OpenCode session analyzer for breaking down context usage per tool and debugging the largest tool calls that wasted the most tokens. Load this skill when you need to inspect OpenCode token usage, identify oversized tool calls, or compare duration breakdowns for slow tools and subagents.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/remorses/contextanalyzer --skill contextanalyzer命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
基于 SOC 职业分类
正在显示 SKILL.md
| name | contextanalyzer |
| repo | https://github.com/remorses/contextanalyzer |
| description | OpenCode session analyzer for breaking down context usage per tool and debugging the largest tool calls that wasted the most tokens. Load this skill when you need to inspect OpenCode token usage, identify oversized tool calls, or compare duration breakdowns for slow tools and subagents. |
contextanalyzer reads OpenCode session history and prints terminal histograms for context tokens and wall-clock duration.
Every time you use contextanalyzer, run the full help first:
npx -y contextanalyzer --help
Never pipe the help output through head, tail, sed, awk, or any other truncation command.
Also fetch and read the latest README in full:
curl -L https://raw.githubusercontent.com/remorses/contextanalyzer/main/README.md
Never truncate the README output. It is the canonical documentation for examples, options, and how token estimates are computed.