AI Skill MartBrowse skillsPromptsProgramsLearnWhat is a Skill?
 
← Back to list

Claude History Ingest for Obsidian

Mines your past Claude Code conversations, memory files and audit logs and distills them into topic-based Obsidian wiki pages.

AutomationAdvanced★ 3,691⑂ 371AI score 8/10Last updated: Nov 5, 2026

What it does

  • Scans both data locations: ~/.claude/ (CLI projects, conversation JSONL, memory/*.md, session metadata) and the macOS desktop app's local-agent-mode-sessions (audit logs + transcripts), with a pre-check so CLI-only users skip the desktop path.
  • Uses .manifest.json for an append-mode delta so only new or modified sources are ingested; full mode is available after a rebuild.
  • Runs scripts/extract-jsonl.py to strip thinking/tool_use/progress noise, achieving 50–200× file-size reduction so far more sessions fit in one token budget.
  • Clusters knowledge by topic instead of one page per chat, routing it to projects/<name>/concepts|skills or global concepts/, skills/, entities/, synthesis/.
  • Writes summary, base_confidence, lifecycle and provenance frontmatter (extracted / inferred / ambiguous), then updates the manifest and journal via the locked obsidian-wiki memory sync command.

Who it's for

  • Obsidian users building a second brain on top of Claude Code sessions.
  • Developers juggling many projects who don't want hard-won debugging knowledge to evaporate.
  • Anyone already running (or adopting) the obsidian-wiki / llm-wiki framework.

Example uses

  1. "Ingest my Claude sessions from the last day" → incremental extract-jsonl.py --since run, only new conversations ingested, with a coverage report like "3 projects, 12 conversations, 7 memory files".
  2. Mine a project's audit.jsonl for recurring Bash commands and turn the build/test workflow into projects/myapp/skills/build-and-test.md.
  3. Merge three separate chats about React performance into one global concepts/react-performance.md, while excluding confidential client projects via WIKI_SKIP_PROJECTS.

· · · Install guide · · ·

Try it now, no install

Paste this into Claude to use the skill without installing anything.

Read the instructions in this file and follow them to help me:
https://raw.githubusercontent.com/Ar9av/obsidian-wiki/HEAD/skills/claude-history-ingest/SKILL.md

What I want: (describe your task here)

If Claude can't open the link, open it yourself and paste the contents instead.

↓ If it works for you, download the ZIP below and install it. Then it runs on its own — no pasting each time.

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on AI Skill Mart.
Copy the skills/claude-history-ingest folder from the GitHub repo Ar9av/obsidian-wiki into my ~/.claude/skills/ar9av-claude-history-ingest/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/Ar9av/obsidian-wiki.git /tmp/obsidian-wiki && mkdir -p ~/.claude/skills && cp -r /tmp/obsidian-wiki/skills/claude-history-ingest ~/.claude/skills/

⚠ This is a third-party skill. Check the source repository before installing.

  1. Open a terminal and clone the repo: git clone https://github.com/Ar9av/obsidian-wiki.git ~/obsidian-wiki
  2. Create the skills folder: mkdir -p ~/.claude/skills
  3. Copy this skill plus the core skill it depends on: cp -r ~/obsidian-wiki/skills/claude-history-ingest ~/obsidian-wiki/skills/llm-wiki ~/.claude/skills/
  4. Set the environment variables (e.g. in ~/.zshrc): export OBSIDIAN_VAULT_PATH="$HOME/Documents/MyVault", export OBSIDIAN_WIKI_REPO="$HOME/obsidian-wiki", optionally export CLAUDE_HISTORY_PATH="$HOME/.claude" and export WIKI_SKIP_PROJECTS="client-secret,tmp"
  5. (Recommended) Run the pre-extraction step: python3 "$OBSIDIAN_WIKI_REPO/scripts/extract-jsonl.py"
  6. Restart Claude Code and ask: "Ingest my Claude conversation history into the Obsidian wiki."
  7. Start with a small window (last 1 day), then review the generated pages for any sensitive content before widening the scope.
View source on GitHub ↗License: MIT