- Introduced tools/heal.py to automatically dynamically identify and build missing structural concepts and entities by tracing contextual usages using litellm.
- Add docs/automated-sync.md with cron/launchd orchestration best-practices.
- Closes Issue #16 on Graph Integrity Constraints.
- Lead with one-sentence hook + output structure upfront
- Add What You Get section naming concrete deliverables
- Consolidate agent compatibility into schema file table
- Add tech stack one-liner
- Streamline use cases, quick start, and graph sections
Each use case now shows exactly what to ingest, what to query, and
what the wiki produces — Research, Reading, Personal KB, Business
intelligence, and Competitive analysis.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Deleted client/, server/, and steps_to_run.md — remnants of the
original GPT-Agent project. Repo is now purely the LLM Wiki Agent.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The wiki now works with any coding agent that reads a repo config file.
CLAUDE.md → Claude Code, AGENTS.md → Codex/OpenCode, GEMINI.md → Gemini CLI.
README updated with agent compatibility table.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Add .claude/commands/ slash commands (/wiki-ingest, /wiki-query,
/wiki-lint, /wiki-graph) so the wiki works directly inside Claude Code
without any Python or API key. CLAUDE.md updated to lead with Claude
Code usage. Python tools remain as optional standalone scripts.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Replaces dual-agent demo with a full personal knowledge base system
where Claude reads source documents and incrementally builds and
maintains a structured, interlinked wiki of markdown pages.
- tools/ingest.py: reads a source, extracts knowledge, updates wiki pages
- tools/query.py: queries the wiki with Claude, optionally files answers back
- tools/lint.py: health-checks the wiki (orphans, contradictions, gaps)
- tools/build_graph.py: two-pass graph builder (wikilinks + Claude inference)
with Louvain community detection and vis.js interactive HTML output
- CLAUDE.md: schema and workflow instructions for the LLM
- wiki/: starter index, log, and overview pages
- raw/, graph/: directory scaffolding
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add .env.example with required environment variables
- Add helpful warning message when google_client_id/secret are missing
- Update steps_to_run.md with detailed setup instructions
- Include instructions for creating Google OAuth credentials
Addresses #9
- Add GitHub stars badge and other relevant badges
- Improve structure with clear sections
- Add features list with descriptions
- Include installation/quick start instructions
- Add tech stack and related projects
- Better formatting and readability
