Add read-only external workspace scanner

[sigilabcodex]

Purpose

Add a read-only scanner that accepts an external root path and returns a stable inventory of literary/editorial project structures without requiring migration into TRURL conventions.

This is the core prerequisite for opening real workspaces such as /media/diegom/luna/DEV/0w, /0w/fiction, and /0w/d-503.

Scanner should detect

• Git repositories and nested Git repositories.
• Markdown files.
• metadata.yaml files.
• README.md files.
• Known folder roles, such as manuscript, drafts, final, working, canon, planning, research, assets, images, derived, dist, published, prompts.
• Collections and child stories.
• Essay/project folders.
• Derived outputs and export artifacts.
• Warnings for ambiguous or risky structures.

Initial targets

The scanner should eventually handle:

• /media/diegom/luna/DEV/0w/fiction
• /media/diegom/luna/DEV/0w/d-503

Examples include:

• fiction/horror/los-andares-del-diablo/stories/*/metadata.yaml
• nested repos under story projects
• D-503 essays with metadata.yaml, final/, working/, derived/, and published/

Acceptance criteria

• Scanner runs in read-only mode.
• Scanner accepts a root path.
• Scanner produces stable JSON.
• Scanner detects nested Git repositories.
• Scanner detects metadata.yaml files.
• Scanner identifies known folder roles.
• Scanner distinguishes source-like folders from derived/dist folders.
• Tests cover fixture folders resembling /0w/fiction and /0w/d-503.
• No write operations are performed during scanning.

Design principle

First let TRURL say: “this is what I found, this is what I understand, this is what I do not understand yet.” Do not add agent edits or Git write actions before the scanner is reliable.