lexicon init

Usage

lexicon init

Description

Initializes lexicon in the current repository by creating the directory structure and manifest file. This is an interactive workflow that prompts for project metadata.

The command discovers the repository root by walking up from the current directory looking for a .lexicon/ directory or Cargo.toml. It falls back to the current directory if neither is found.

If lexicon has already been initialized (.lexicon/manifest.toml exists), the command fails with an error.

What It Creates

.lexicon/ directory

Path	Purpose
.lexicon/manifest.toml	Root manifest with project metadata
.lexicon/context/	Assembled context for AI consumption
.lexicon/conversations/	Recorded conversation sessions
.lexicon/audit/	Audit trail records
.lexicon/cache/	Cached computation results
.lexicon/state/	Runtime state

specs/ directory

Path	Purpose
specs/contracts/	Contract TOML files
specs/behavior/	Behavior scenario definitions
specs/scoring/	Scoring model
specs/conformance/	Conformance specifications
specs/non_goals/	Non-goal definitions

Interactive Prompts

The init workflow asks for:

1. Project name — Used as the project identifier
2. Description — Short project description
3. Repository type — library, binary, or workspace
4. Domain — What the project does (e.g., “key-value store”, “parser”)

Manifest

The generated manifest.toml includes:

• Project metadata (name, description, type, domain, timestamps)
• Default preferences (kebab-case naming, trait-based conformance)
• Default AI policy (source files editable, specs require review, manifest/gates protected)

Example

$ lexicon init
Initializing lexicon
? Project name: my-parser
? Description: A streaming JSON parser
? Repository type: library
? Domain: parser
Lexicon initialized successfully