Skip to main content

Glossary

A dictionary of terms we use day to day. When someone asks "what is X?" in a DM, just send them the link from here.

It's split into two parts:

  • Our terms — things specific to how we operate (ADR, Bot, Runbook, Eigenoid).
  • General (AI) — generic AI domain concepts that everyone should understand.

Our terms​

  • ADR — Architecture Decision Record
  • Bot — GitHub App or automation
  • Eigenoid — the Python package and the organization
  • Runbook — reproducible operational procedure

General (AI)​

  • Agents — autonomous AI systems that execute tasks
  • Agentic Workflows — multi-step patterns with planning, execution, and verification
  • Context — the token window the model sees
  • Context-Aware AI — AI that adapts its output to the actual project state
  • Grounding — connecting model output to verified data
  • Hallucination — when the model generates incorrect information with confidence
  • LLM — Large Language Model
  • MCP — Model Context Protocol
  • Prompts — the instructions that guide the model
  • RAG — Retrieval-Augmented Generation
  • SDD — Spec-Driven Development
  • Skills — domain knowledge packages for agents
  • Tokens — the units LLMs process
  • Tool Use — how agents invoke external tools

How to add a term​

Each term is a file in docs/glossary/<term>.md with minimal frontmatter:

---
term: Name
aliases: [variant1, variant2]
---
yaml

A short 1-3 paragraph definition. If it needs more than a page, it's probably not a glossary term — turn it into an ADR, runbook, or dedicated doc instead.