FairladyZ625/conversation-mining

آخر نشاط مصدر مسجل: ٢٣ يونيو ٢٠٢٦ في ١٤:٤٦

آخر تحديث لفهرس SkillsMP: ١٢ نوفمبر ٢٠٢٦ في ١٢:٥٢

skills مجمعة: ١

نجوم GitHub: ٣
تفرعات GitHub: ١

التثبيت بمساعدة مساعد ذكي

انسخ هذا الطلب إلى المساعد الذكي الذي تستخدمه.

Follow https://skillsmp.com/skill-install/prompt.md to install Agent Skills from https://github.com/FairladyZ625/conversation-mining.

ملف README للمستودع

مصدر ملف README للمستودع: README.md

قراءة الملف الأصلي

عرض ملف README كاملًا

# conversation-mining [![skills.sh](https://skills.sh/b/FairladyZ625/conversation-mining)](https://skills.sh/FairladyZ625/conversation-mining) Export local Claude Code, Codex, and Antigravity conversations into Markdown plus a static HTML viewer. This project is designed for local forensics and recall: - backfill or incrementally export conversation history - browse everything in one viewer - deep-link into a specific conversation - copy a conversation reference/prompt for handoff to another AI - inspect AG task artifacts when Antigravity leaves local files behind It can be used in two ways: - as a standalone local script/tool - as an installable skill via [SKILL.md](./SKILL.md) ## Copy This To Your AI ```text Install and learn `conversation-mining` from: https://github.com/FairladyZ625/conversation-mining If supported, install the skill with: npx skills add FairladyZ625/conversation-mining --skill conversation-mining Read README.md, SKILL.md, and AI_INSTALL.md. Install it locally as both a skill and a CLI if supported. Verify with: conversation-mining --no-open --days 1 ``` ## What It Supports - Claude Code - reads local `~/.claude/projects/*/*.jsonl` - reconstructs mainline conversations and embedded sub-agent relationships - Codex - reads local `~/.codex/sessions`, `history.jsonl`, and `state_5.sqlite` - Antigravity - reads local `~/Library/Application Support/Antigravity/.../state.vscdb` - extracts trajectory summaries and links AG brain artifacts from `~/.gemini/antigravity/brain/<uuid>/` when available ## Current Status This is a pragmatic local-export tool, not an official SDK. - Claude and Codex usually provide much richer transcript coverage - Antigravity currently behaves more like task-summary extraction than full raw transcript recovery - AG artifact linking works when a stable local brain workspace can be matched ## Quick Start ```bash git clone git@github.com:FairladyZ625/conversation-mining.git cd conversation-mining python3 convs.py ``` That will: 1. export recent conversations 2. write JSON + viewer assets into `./exported_conversations` 3. write human-readable Markdown transcripts into a separate transcript directory 4. build `./exported_conversations/index.html` 5. open the viewer on macOS ## Install As a CLI Editable install: ```bash python3 -m pip install -e . conversation-mining --no-open ``` Run as a module: ```bash python3 -m conversation_mining --days 7 ``` With `pipx`: ```bash pipx install . conversation-mining --date 2026-03-19 ``` ## Use As a Skill This repository also includes [SKILL.md](./SKILL.md). It can be installed with the open Agent Skills CLI: Preview the available skill: ```bash npx skills add FairladyZ625/conversation-mining --list ``` Install into the current project: ```bash npx skills add FairladyZ625/conversation-mining --skill conversation-mining ``` Install globally for Codex: ```bash npx skills add FairladyZ625/conversation-mining \ --skill conversation-mining \ --agent codex \ --global \ -y ``` Verify a global Codex install: ```bash npx skills list --global --agent codex ``` If your agent supports skill installation from a git repository, install this repo as a skill and then invoke `conversation-mining` for: - locating a conversation by date or id - exporting recent history - opening the local viewer - generating a handoff reference/prompt ## Common Commands Export today and build the viewer: ```bash python3 convs.py --no-open ``` Backfill the last 7 days: ```bash python3 convs.py --days 7 ``` Export a specific date: ```bash python3 convs.py --date 2026-03-19 ``` Write to a custom output directory: ```bash python3 convs.py --output-dir ~/tmp/conversation-mining-output ``` Write Markdown transcripts to a custom directory: ```bash python3 convs.py --markdown-dir /Users/lizeyu/Documents/ZeYu-AI-Brain/LOCAL-LARGE-FILES/agent-context/conversation-transcripts ``` Advanced: export only JSON/Markdown without building the viewer: ```bash python3 export_all.py --days 7 --output-dir ./exported_conversations ``` ## Output Structure ```text exported_conversations/ conversations.json index.html ``` Separate human-readable transcripts: ```text /Users/lizeyu/Documents/ZeYu-AI-Brain/LOCAL-LARGE-FILES/agent-context/conversation-transcripts/ 2026-03-19/ claude_*.md codex_*.md ag_*.md ``` ## Viewer Features - source filters: Claude / Codex / AG - search by title, message text, workspace, tags, and AG artifact metadata - light / dark mode - reading modes: - mainline - mainline + sub-agents - forensics - quick summary cards: - final answer - key files - commands - signals - timeline replay - bookmarks, tags, collections - workspace grouping - related task thread navigation - AG artifact cards with direct local file links ## Known Limits - AG is not guaranteed to have a full chat transcript store on disk - AG artifact coverage is partial and depends on stable local brain workspace ids - the viewer is optimized for local file usage - localStorage is used for bookmarks/tags/collections ## Privacy This tool reads from local application state and local conversation history. Before publishing exported data, review it carefully: - conversation content may contain secrets - copied transcript Markdown can include absolute file paths - AG artifact paths may expose local workspace layout ## License MIT

جارٍ تحميل ملف README للمستودع…

Skills في هذا المستودع

المهن الحاسوبية الأخرى

١ فئات مهنية · 100% مصنفة

skill

المهنة

الوصف

آخر تحديث

conversation-mining

المهن الحاسوبية الأخرى

统一检索和使用 Claude Code / Codex / AntiGravity 的历史对话。支持按日期、关键词、会话 ID、Conversation Reference、Conversation Prompt 找回原文，也支持在用户明确要求时，对指定会话或某天会话做提炼、总结、笔记整理。

لغة النص الأصلي: الصينية

2026-06-23

عرض ١ من أصل ١ skills مجمعة.