scholaraio

Repository Knowledge Design

Status: Current map

Last Updated: 2026-06-12

ScholarAIO is an agent-first research infrastructure. Its published repository knowledge is organized for progressive disclosure: entry docs give agents a map, and the public docs/ tree carries user-facing and durable design material.

This layout is inspired by OpenAI’s harness-engineering guidance that AGENTS.md should behave like a table of contents rather than a monolithic manual: https://openai.com/index/harness-engineering/

Principles

Directory Roles

Path Role
docs/index.md Public documentation home
docs/getting-started/ User setup and upgrade paths
docs/guide/ User-facing workflow references
docs/writing-guide/ Writing and diagram authoring guides
docs/api/ Python API documentation
docs/design-docs/ Long-lived architecture and runtime design decisions
docs/product-specs/ Product behavior and workflow specifications
docs/generated/ Generated facts such as schema or CLI snapshots
docs/superpowers/ Local planning artifacts created by superpowers workflows
docs/archive/ and docs/internal/ Historical or local-only notes excluded from published docs

Reading Order

  1. Start with AGENTS.md for hard constraints and navigation.
  2. Use this file for the repository knowledge map.
  3. Use docs/guide/agent-reference.md for agent runtime and skill details.
  4. Use the relevant design doc, product spec, generated reference, or guide before changing a governed area.
  5. Use source code and tests as the final authority for current behavior.

Maintenance Rules