A Claude skill that designs any process, idea, or problem into an ICM workspace — folder structure as agent architecture — or restructures an existing folder, repo, or vault into one.
ICM (Interpretable Context Methodology) replaces orchestration code with structure: numbered folders carry sequencing, hierarchy carries context scoping, plain markdown files carry state. One agent, reading the right files at the right moment, does the work of a multi-agent framework — and a human can open any folder and see exactly what state the system is in.
The workspace is a library. The routing files are the catalog: small, stable, they point at everything and store almost nothing. One librarian — one model — walks the building, and the question decides which shelf gets walked to.
- Paper: Interpretable Context Methodology: Folder Structure as Agent Architecture (Van Clief & McDermott)
- Community: Clief Notes
Two modes:
- Build — extracts the structure already present in how you describe your work (the stages, the human gates, what's stable vs. per-run), picks one of five proven forms, and scaffolds the smallest workspace that carries it.
- Restructure — audits an existing folder, classifies every file (catalog / contract / factory / product / dead), proposes a migration map for approval, then migrates and validates.
Five forms, one skeleton: Pipeline (production line), Umbrella (portfolio of pipelines), Record library (people/clients/sessions), Knowledge bundle (a navigable brain), Context map (an organization as a graph). They compose and recurse.
Every result is validated with the walk test: an agent with no memory must orient, act, and report status from the files alone.
Claude Code: copy this folder to ~/.claude/skills/icm-architect/ (or .claude/skills/icm-architect/ inside a project), then ask Claude to "ICM this" / "structure this for agents" / "build me a workspace for X".
Claude apps: upload icm-architect.skill (build it with the skill-creator packager, or zip this folder's contents) via Settings → Capabilities.
icm-architect/
├─ SKILL.md the method: invariants, build mode, restructure mode, walk test
├─ references/
│ ├─ core.md five principles, five-layer hierarchy, naming, token discipline
│ └─ forms.md the five forms in depth: skeletons, moves, failure modes
└─ assets/templates/ copyable starters: CLAUDE.md, CONTEXT.md, stage contract,
node card, schema, setup questionnaire
MIT licensed, like the protocol it serves.