Help us improve
Share bugs, ideas, or general feedback.
Share bugs, ideas, or general feedback.
Share bugs, ideas, or general feedback.
By sashabogi
Unified codebase intelligence, semantic memory, and multi-provider LLM routing
npx claudepluginhub sashabogi/foundation-plugin --plugin foundationShow unified memory statistics from Gaia and Open Brain. Displays total memories, tier breakdown, top topics, and more. Trigger: /foundation:brain-stats
Seldon — reconcile a project's documentation against reality: classify every doc (category/status/boundary), resolve version/supersession clusters, and optionally scan the code to find work that's BUILT but undocumented (and docs describing unbuilt features). Trigger: /foundation:doc-inventory
Seldon — generate a browsable HTML documentation portal from a project's markdown docs. Markdown stays the source of truth; the portal is generated (never hand-maintained), styled with the project's design system, and served locally. Run after /foundation:doc-restructure. Trigger: /foundation:doc-portal
Seldon — restructure a project's docs into a standard taxonomy and build the hierarchical AGENTS.md context tree. Moves docs into typed folders, archives superseded ones, generates an INDEX, and authors root + per-boundary AGENTS.md. Run /foundation:doc-inventory first. Trigger: /foundation:doc-restructure
Launch the Foundation dashboard UI in the browser. Starts the local server on port 3333 showing Demerzel snapshots, Gaia memories, and session history. Trigger: /foundation:foundation-ui
Matches all tools
Hooks run on every tool call, not just specific ones
Admin access level
Server config contains admin-level keywords
Executes bash commands
Share bugs, ideas, or general feedback.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge.
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge.
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Optimized file search, semantic indexing, and persistent memory for Claude Code — with optional sync to a self-hosted web dashboard
Cognitive layer for Claude Code — code navigation, conversation recall, safety hooks, behavioral learning
Persistent memory system for AI coding sessions — cross-tool memory sharing with 6-dimensional hybrid search
Persistent project memory for AI coding agents — semantic search, AST-aware chunking, dependency graphs, and conversation history
Persistent memory across Claude Code sessions using Cognis
CodeAlive context engine for semantic code search and AI-powered codebase Q&A. Enables AI coding agents to understand entire codebases beyond just open files — search across all indexed repositories, trace cross-service dependencies, discover usage patterns, and get synthesized answers to architectural questions. Includes a lightweight code exploration subagent, authentication hooks, and multiple search modes (fast lexical, semantic, and deep cross-cutting). Works standalone or alongside the CodeAlive MCP server for direct tool access via the Model Context Protocol.
Trantor — the hub-world for AI agent crews: live message bus, presence, project Kanban/flow board + crew orchestration for independent AI coding agents (Claude, Codex, Gemini, Kimi, DeepSeek)
Hook triggers when Bash tool is used
Hook triggers when Bash tool is used
Requires secrets
Needs API keys or credentials to function
Requires secrets
Needs API keys or credentials to function
Share bugs, ideas, or general feedback.
"The future is not set, but it can be guided." -- Hari Seldon

Codebase intelligence and semantic memory for Claude Code, inspired by Asimov's Foundation universe.
Foundation is a Claude Code plugin that wraps the @sashabogi/foundation npm package and gives your AI assistant persistent memory, deep codebase understanding, and automatic session lifecycle management. The plugin itself is a thin shell: it contributes hooks that fire on lifecycle events at zero token cost, skills that load instructions only when you invoke them, a local UI server, and a stable adapter to the bundled MCP server. The MCP server (shipped by the npm package) registers 29 tools spanning Demerzel (codebase intelligence), Gaia (local memory), and a small provider registry (provider_list / provider_test). In Claude Code 2.1+ those tools are loaded on demand via the deferred-tool system, so they cost no context until you call them.
The naming comes from Isaac Asimov's Foundation series. In the novels, Hari Seldon created the Foundation to preserve human knowledge through the collapse of the Galactic Empire -- a millennia-long dark age where everything would otherwise be forgotten. This plugin does the same thing for your development work: it preserves architectural decisions, codebase understanding, and project context across sessions, projects, and tools, so nothing is lost when a conversation ends or a context window resets.
Three core systems power the plugin:
/foundation:doc-inventory, /foundation:doc-restructure, /foundation:doc-portal) plus a SessionStart hook that auto-loads the nearest AGENTS.md context tree. (Seldon was previously Foundation's multi-agent orchestrator; that role has been sunsetted -- orchestration now lives in Scrooge and Claude Code's native agents -- and the name was repurposed for the documentation pillar.) See the Seldon section below for details.# Add the marketplace
claude plugin marketplace add sashabogi/foundation-plugin
# Install the plugin
claude plugin install foundation
# Restart Claude Code
That's it. No environment variables required for core functionality. Open Brain cloud memory is optional (see setup below).
"I have been watching for 20,000 years."
Demerzel creates a structural understanding of your codebase without requiring Claude to read every file. When you generate a snapshot (via /foundation:snapshot), Demerzel walks your project tree, parses imports and exports, builds a symbol index, maps the dependency graph, and writes a compact representation to .foundation/snapshot.txt.
From that snapshot, Claude can:
The snapshot captures: file paths, line counts, import/export relationships, symbol locations, and the full dependency graph (both import graph and reverse export graph). Search operations against the snapshot are free -- they cost zero tokens beyond the results themselves.
