copilot-history-ingest

Extract knowledge from GitHub Copilot CLI session history into Obsidian wiki pages.

Updated Nov 9, 2023
One-click install
npx skills add https://github.com/oresttokovenko/dot-files --skill copilot-history-ingest-oresttokovenko
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: copilot-history-ingest
Source: https://github.com/oresttokovenko/dot-files/tree/main/chezmoi/dot_agents/skills/copilot-history-ingest
Command: npx skills add https://github.com/oresttokovenko/dot-files --skill copilot-history-ingest-oresttokovenko

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? GitHub Copilot CLI sessions contain valuable architecture decisions, debugging notes, and patterns, but they sit scattered across session-state directories, SQLite databases, and VS Code workspace storage. This Skill mines those conversations and distills them into searchable, topic-clustered Obsidian wiki pages instead of letting the knowledge stay buried in raw transcripts. ## Core Features & Use Cases - Multi-source ingestion: Scans all three Copilot data locations — per-session directories under ~/.copilot/session-state, the global session-store.db SQLite database, and VS Code workspaceStorage transcripts and memory artifacts. - Delta-aware append mode: Tracks ingested sessions in a .manifest.json file so only new or modified sessions are processed on repeat runs, with a full mode for rebuilds. - Topic clustering and distillation: Groups knowledge by topic across sessions rather than one page per session, routing content to project concepts, skills, entities, or synthesis pages with provenance markers and confidence metadata. - Use Case: After a week of Copilot-assisted debugging and feature work, run the ingest to capture what you learned — the skill surfaces recurring patterns like API rate-limiting fixes into searchable wiki pages linked to the relevant projects. ## Quick Start Ask the assistant to ingest your Copilot CLI sessions into your Obsidian vault, optionally specifying append mode for only new sessions since the last run.

Frequently Asked Questions about copilot-history-ingest

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I ingest GitHub Copilot CLI sessions into Obsidian?

Point the skill at your Copilot data locations and it scans session-state directories, session-store.db, and VS Code transcripts, then distills the content into wiki pages. It resolves your vault path from config and updates a manifest so repeat runs only process new sessions.

Where does GitHub Copilot CLI store session history?

Copilot stores data in three places: per-session directories under ~/.copilot/session-state with events.jsonl and checkpoints, a global SQLite database at ~/.copilot/session-store.db, and VS Code workspaceStorage under GitHub.copilot-chat with transcripts and memory artifacts.

Does the ingest reprocess all sessions every time?

No. Append mode is the default and uses a .manifest.json file at the vault root to skip sessions already ingested, processing only new sessions or ones whose updated_at timestamp changed. A full mode exists for rebuilds after a wiki reset.

What data sources give the highest-value content from Copilot sessions?

Checkpoints are the highest-value source since they contain pre-distilled summaries with overview, work_done, technical_details, and next_steps. Session summaries and user-saved memory artifacts rank next, with raw conversation turns processed selectively.

Does ingesting Copilot history copy private conversation text into the wiki?

No. The skill distills and synthesizes knowledge rather than copying raw conversation text, skips internal reasoning fields entirely, and filters out anything resembling secrets, API keys, or tokens. Sensitive personal content is flagged for user confirmation first.