Project Operating Manual
A project operating manual is a per-project context file, often a CLAUDE.md-style file plus a memory folder, that tells an agent what a project is, what it is trying to achieve, what decisions are already made, and where evolving memory lives [src-059].
Key points
- Roberts recommends creating six to eight project folders that match the user’s real life and business domains [src-059].
- Each folder should include a top-level operating manual explaining the folder purpose, goal, stack, made decisions, memory map, and relevant references [src-059].
- The manual should stay compact enough to be injected into conversations without bloating context; Roberts suggests keeping it under about 200 lines [src-059].
- The manual is mutable: active strategies and next actions can change as the project evolves [src-059].
- This is the mid-term layer of the memory system and connects directly to Codex Project Workspace and Agent Harness Portability [src-059].
Related entities
Related concepts
- Three-Layer AI Memory
- AI Memory Operating System
- Codex Project Workspace
- Agent Harness Portability
- Global vs Project Skills
- Claude Code Context Management Discipline
Source references
- [src-059] Jack Roberts — “This Memory System just 10x’d Claude Code” (2026-05-03)
Recommended next
Keep reading from this thread
From 494 indexed pages and articles.
- Wiki concept Three-Layer AI Memory Roberts's structure for giving an assistant durable context: short-term memory answers "who am I?", mid-term memory answers "what am Related by 059
- Wiki concept Codex Project Workspace A Codex project workspace is the practice of organizing Codex into project folders that mirror real areas of work, then running chats, agents Related by project
- Insight AI Beyond POCs How enterprise AI moves beyond proofs of concept through ownership, governance, measurement, adoption, and production operating models Readers have engaged with this next