This is an early release preview. You may encounter bugs.
Obsidian Mind logo
Unclaimed

Tool coding memory

Obsidian Mind

A self-organizing Obsidian vault that gives AI coding agents persistent memory. Claude Code, Codex CLI, Gemini CLI.

A 85/100 GitHub score ? This grade is derived from GitHub signals, not user votes. Open for the full breakdown.
Discussion

01 / Summary

About Obsidian Mind.

Obsidian Mind is an Obsidian vault template that gives AI coding agents memory across sessions. You open the vault in Claude Code (full support), Codex, or Gemini CLI and talk about your work, and the agent writes, files, and links notes about projects, people, decisions, incidents, and wins, so each session starts from what earlier ones recorded.

The split is deliberate: procedural hook scripts handle classification, validation, indexing, and context injection, while the agent handles writing and linking. A session-start hook injects your goals, active projects, recent changes, open tasks, and the vault file listing. Other hooks classify each message and add routing hints, validate frontmatter and wikilinks after a note is written, back up transcripts before context compaction, and report vault-hygiene drift after each response. On Claude Code, an optional plugin delivers the session context as an instruction file instead of hook output.

Retrieval works best with Qmd, which runs three small local models (embedding, query expansion, and reranking) for hybrid keyword and semantic search, exposed to the agent as MCP tools. Without it the vault falls back to grep and the Obsidian CLI. A separate MCP server lets a coding session in any other repository search, read, and write back to the vault.

The vault also doubles as a performance record: work notes link to competency notes, so the backlinks become evidence for review season, and commands assemble brag documents, review briefs, self-assessments, and peer reviews from what the vault holds.

  • Session lifecycle hooks: five hooks inject context at start, classify each message, validate written notes, back up transcripts, and report drift
  • Slash commands: standup, brain dump, wrap-up, weekly synthesis, 1:1 and incident capture, review briefs, vault audit, and tidy-up
  • Subagents: helpers for spotting uncaptured wins, loading context, cross-linking notes, profiling people from Slack, and fact-checking review drafts
  • Semantic search: optional Qmd integration for local hybrid search, exposed as MCP tools, with grep and Obsidian CLI fallback
  • Cross-repo memory: an MCP server that lets sessions in other repositories search, read, and record to the vault
  • Performance graph: competency notes collect backlinks from work notes as evidence for reviews and brag documents
  • Obsidian Bases dashboards: views for active work, recently touched notes, incidents, people, 1:1 history, and competencies
  • Multi-agent support: full support in Claude Code, with shared hooks and commands for Codex and Gemini CLI
  • Vault migration: imports and classifies content from an existing vault

02 / Discussion CREDIBILITY-GATED

Discussion

Reading is open to everyone. Posting and voting need a verified identity or a GitHub grade of B or higher.

  • No discussions yet.

04 / Build

Build with Obsidian Mind.

Browse the catalog for frameworks, tools, and harnesses, each scored on real GitHub credibility.

Get Obsidian Mind →

Browse the catalog