Token导航 LogoToken导航TokenDH.com
开发只读github未标认证来源可访问许可证需确认审计提醒

mega-code-wisdom-gen超级代码智慧生成器

Agent Skill

mega-code-wisdom-gen 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

343

周安装

14

GitHub Stars

43

下载量

111
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:mega-code-wisdom-gen(超级代码智慧生成器)
来源仓库:https://github.com/wisdomgraph/mega-code
仓库路径:skills/mega-code-wisdom-gen
安装命令:
npx skills add https://github.com/wisdomgraph/mega-code --skill mega-code-wisdom-gen
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/wisdomgraph/mega-code --skill mega-code-wisdom-gen

简介

mega-code-wisdom-gen 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息。

  • 适合在 Codex、Claude、Cursor、Gemini CLI 中围绕仓库状态、代码变更或协作事项进行整理。
  • 可协助分析代码提交历史和协作流程中的关键节点。
  • 安装前需确认权限范围和维护状态,注意是否涉及联网或文件读写。
  • 建议结合原始 README 进一步了解具体用法和功能边界。

SKILL.md

Run Skill Extraction Pipeline

Extract reusable skills and coding strategies from your Claude Code sessions.

⚠️ Important: Pipeline is Long-Running

The pipeline command blocks until the server finishes processing. The server runs the pipeline asynchronously and this client polls for completion.

DO NOT interrupt the command (Ctrl+C) while it is running. Silence with no new output for several minutes is completely normal during LLM inference (especially with gpt-5-mini or other reasoning models).

The default poll timeout is 20 minutes. For longer runs, use --poll-timeout:

  • --poll-timeout 3600 — wait up to 1 hour
  • --poll-timeout 0 — wait indefinitely (no timeout)

Setup

MEGA_DIR="$(cd "${CLAUDE_SKILL_DIR}/../.." && pwd)"
uv run --directory "$MEGA_DIR" python -m mega_code.client.check_auth

If the auth check fails (non-zero exit), show the output to the user and stop.

All commands below assume MEGA_DIR is set.

Flags

FlagBehavior
*(none)*Process current session only
--projectAll sessions in current project
--project @nameSpecific project by name prefix
--session-id <uuid>Specific session
--model <alias>LLM model (default: server picks best)
--poll-timeout <seconds>Max seconds to poll for completion (default: 1200 = 20 min; 0 = indefinite)
--include-claudeInclude related Claude Code sessions from the project

Project argument formats (all equivalent): @mega-code · mega-code · mega-code_b39e0992 · /path/to/project

Running the Pipeline

All variables must be in one single Bash call so $LOG and $MEGA_DIR stay in scope:

LOG="/tmp/mega-code-run-$(date +%Y%m%d-%H%M%S).log" && \
  echo "Pipeline log: $LOG" && \
  export CLAUDE_PROJECT_DIR="$PWD" && \
  uv run --directory "$MEGA_DIR" python -m mega_code.client.run_pipeline [FLAGS] 2>&1 | tee "$LOG"

Replace [FLAGS] with desired combination from the table above. Tell the user the log path so they can monitor with tail -f or check after completion.

Model Options

AliasProvider
gemini-3-flashGoogle
gpt-5-miniOpenAI

When omitted, server selects based on configured LLM keys (priority: Gemini > OpenAI). Falls back to gemini-3-flash.

Pipeline Outputs

  1. Skills & Strategies — saved to pending dirs for review/install
  2. Lesson Learned documents — saved to ~/.local/share/mega-code/data/feedback/{project_id}/{run_id}/lessons/ (from sessions tagged lesson_learn)

Handling Active Pipeline (Exit Code 2)

If the pipeline command exits with code 2, a pipeline is already running. Parse the JSON output to get conflict.run_id and conflict.project_id.

Use the AskUserQuestion tool to present these options:

Question: "A pipeline is already running for this project (run_id: {run_id}). What would you like to do?"

Options:

  1. "Stop it and start a new one"
  2. "Wait for the existing run to finish"
  3. "Leave it running — exit without action"

Option 1 — Stop and restart:

uv run --directory "$MEGA_DIR" python -m mega_code.client.cli pipeline-stop --run-id <RUN_ID>

Then re-run the pipeline command from "Running the Pipeline" section.

Option 2 — Wait for existing run:

uv run --directory "$MEGA_DIR" python -m mega_code.client.run_pipeline \
  --poll-existing <RUN_ID> --project <PROJECT_ID> [--poll-timeout <seconds>] 2>&1 | tee "$LOG"

Then follow the Post-Pipeline Workflow as normal.

Option 3 — Leave it running: Return immediately. Do not print anything or ask further questions.

Handling Server Timeout (Exit Code 3)

If the pipeline command exits with code 3, the pipeline exceeded the server's max runtime and was terminated. Parse the JSON output for timeout.error details.

Use the AskUserQuestion tool to present these options:

Question: "The pipeline timed out on the server ({error message}). What would you like to do?"

Options:

  1. "Run again — start a fresh pipeline run"
  2. "Do nothing — exit without action"

Option 1 — Run again: Re-execute the pipeline command from "Running the Pipeline" section.

Option 2 — Do nothing: Return immediately. Do not print anything or ask further questions.

Post-Pipeline Workflow (MANDATORY)

The pipeline prints a JSON object with additionalContext on completion. You MUST parse and follow the embedded workflow immediately — do NOT just report "pipeline complete".

Steps:

  1. Parse run_id and project_id from the pipeline output JSON (additionalContext).
  2. Run this command to get the detailed review workflow instructions:
uv run --directory "$MEGA_DIR" python -m mega_code.client.pending review \
  --run-id <RUN_ID> --project-id <PROJECT_ID>
  1. Follow the printed instructions exactly for the review, install, and archive steps. Those instructions are a sub-workflow, not the end of wisdom-gen. When the sub-workflow completes — whether you installed items, archived them, or skipped everything — you MUST return here and continue with the "MANDATORY — Enhance Generated Skills (post-review handoff)" section below. Do NOT terminate the wisdom-gen workflow at the archive step.

MANDATORY — Enhance Generated Skills (post-review handoff)

After the review sub-workflow returns, you MUST run the enhance handoff whenever this pipeline run generated any skills. Install/archive status is irrelevant — an archived skill is still a generated skill and is still eligible. The only valid skip is a run that produced zero skill candidates (strategies/lessons-only). If unsure, default to running the handoff.

"Run the handoff" means execute the trigger check and (when applicable) the binary prompt defined in the reference — *not* "always perform an enhancement". A user "No" answer, or a non-interactive default-to-No, is a valid completion of the handoff, not a skip.

You MUST now read references/enhance-handoff.md and follow it end-to-end before terminating wisdom-gen. It owns the trigger check, the binary Yes/No prompt, and the per-skill enhancement flow. Do not re-implement the decision logic here.

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

36.43%
按下载量换算40

Claude

26.02%
按下载量换算29

Cursor

19.31%
按下载量换算21

Gemini CLI

8.25%
按下载量换算9

安全审计

Gen Agent Trust Hub

通过

Socket

可疑

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills