Workflow Objects

An Obsidian plugin for "typed object" note workflows.

This plugin serves to implement the metadata schemas associated with notes of particular file classes, as defined and managed using Metadata Menu, to curate typed-note file names, locations, and frontmatter.

Relationship to Metadata Menu

Workflow Objects is designed as a companion to Metadata Menu. The two plugins divide responsibilities cleanly:

Metadata Menu defines what an article object is — that is, a note with a metadata-defined schema of the article type: which fields it has, what types they are, what their defaults should be. Workflow Objects implements that definition across the vault: it creates new article objects (notes conforming to the article type schema) pre-populated with those fields, files them in the right folder with the right filename, keeps their frontmatter tidy as the schema evolves, lets you navigate sequentially between article objects, and generates an Obsidian Bases .base file that serves as a live catalog of every article object in the vault.

Together, the two plugins transform a vault of loosely structured notes into a vault of typed objects — Metadata Menu as the schema manager, Workflow Objects as the curator that implements those schemas in practice.

If Metadata Menu is installed, Workflow Objects auto-detects its type-field alias (fileClass by default) and reads type definitions directly from the Metadata Menu fileClass folder — no duplicate configuration needed. It can also operate without Metadata Menu, reading type definitions from any folder of notes you designate.

Installation

1. In Obsidian, open Settings → Community plugins → Browse.
2. Search for Workflow Objects and install.
3. Enable the plugin.

Manual installation: download main.js, manifest.json, and styles.css from the latest release and copy them to .obsidian/plugins/workflow-objects/ inside your vault.

Type definitions

The plugin reads type schemas from a folder of Markdown notes. Each note's filename is the type name — article.md defines the article type, project.md defines the project type, and so on. The note's frontmatter carries the field schema in Metadata Menu's fileClass format:

---
fields:
  - name: author
    type: Input
  - name: status
    type: Select
    options:
      valuesList:
        draft: ""
        published: ""
  - name: rating
    type: Number
---

If Metadata Menu is installed, this folder is read automatically from its settings. Otherwise, set Types path in the plugin settings.

Commands

All commands are available from the Command Palette (Ctrl/Cmd+P).

Create new workflow object

Prompts for a type, then a title. Creates a new typed note — that is, a new object of the chosen type — pre-populated with all fields defined in that type's schema (with their default values). The note is filed at the path and given the filename determined by your path mappings and filename mappings settings.

Navigate to next / previous workflow object of same type

From any typed note (object), jumps to the next or previous note of the same type, sorted by filename. Wraps around at the boundary unless wrap-around navigation is disabled in settings.

Useful for paging through all objects of a given type in sequence — all your book objects (notes of the book type), all your project objects, and so on.

Open workflow object by type

Two-step picker: first choose a type, then choose a note from the list of all objects of that type (all notes whose type field matches). The list is sorted by filename, newest-first by default when filenames carry date prefixes.

Reshelve workflow object

Moves and renames the active note to match its current frontmatter — applying your path and filename template settings to its type and title fields. Use this after changing a note's type or title, or after updating your path mapping rules, to bring the object into its correct location.

If the filename already encodes a date prefix (e.g. 20240315T0930--my-note.md), that prefix is preserved unless the frontmatter contains a more authoritative date field (see Date prefix source fields in settings).

Sort frontmatter

Reorders the active note's frontmatter fields using the default sort mode from settings (schema order or alphabetical). Equivalent to running Sort frontmatter (schema order) or Sort frontmatter (alphabetical) depending on your Default field sort setting.

Sort frontmatter (schema order)

Reorders frontmatter fields to match the sequence defined in the type schema. Fields not present in the schema are moved to the end.

Sort frontmatter (alphabetical)

Reorders frontmatter fields A–Z, regardless of schema.

Clean frontmatter (interactive)

Opens a menu of cleanup actions for the active typed note (object). Each action is a preset combination of three independent operations: adding fields missing from the schema, removing fields not present in the schema, and sorting the result.

When an enforce action would delete fields that currently have values, the plugin shows a confirmation dialog listing exactly which fields and values would be lost before proceeding.

Clean frontmatter (standard)

Non-interactive ensure: adds any fields missing from the type schema and preserves all existing fields — including those not in the schema — using the default sort from settings. Safe to run on any typed note (object).

Clean frontmatter (strict)

Non-interactive enforce: adds missing fields and removes any fields not defined in the type schema, using the default sort from settings. Use with care — there is no confirmation dialog.

Create new workflow object catalog

Generates an Obsidian Bases .base file that serves as a live catalog or index of all objects of a chosen type — that is, all notes whose type field matches the chosen type.

Flow:

1. Choose a type from the suggester.
2. Review or edit the destination directory and filename (pre-filled from your catalog settings).
3. Click Create.

The generated catalog (.base file) includes:

• A filter that selects all notes of the chosen type (all objects whose type field equals the chosen value).
• A Title column rendered as a clickable link using each object's title field.
• One column per schema field. Fields of type File (linked notes — references to other objects) are rendered as titled links to the target object; all other fields are shown as raw property values.
• Optional Date created and Date modified columns (controlled by settings).
• Sentence-case display names for all columns (e.g. workflow-effort → Workflow effort), unless remapping is disabled.

If a .base file already exists at the target path, a collision dialog offers three options: Replace (overwrite), Rename (enter a new filename), or Cancel.

Clone object type

Creates a new object type (fileClass) as a copy of an existing one.

Flow:

1. Choose a source type from the suggester.
2. A dialog opens showing the new type's name (pre-filled as <source>-copy) and all fields from the source schema.
3. Edit the name — collisions with existing type names are caught and flagged live.
4. In the field list, for each field:
   • Uncheck the checkbox to omit the field from the clone entirely.
   • Edit the name input to rename the field in the clone (leave blank to keep the source name).
   • Drag the row or use the ↑ / ↓ buttons to reorder fields.
5. Click Clone.

The new fileClass note is created in the same folder as your other type definitions (the Metadata Menu fileClass folder, or your configured Types path). It is populated with only the fields you kept, under the names and in the order you specified. Each field receives a fresh unique ID so Metadata Menu can track it independently of the source type. Field type, options, and nesting path are preserved from the source.

The new type definition file opens automatically after creation.

Note: The Metadata Menu public API covers field-value operations only and does not expose fileClass creation. The clone is written directly to the vault using the same processFrontMatter mechanism Metadata Menu itself uses to maintain fileClass definitions.

Copy fields from…

Copies a selection of field definitions from one object type into another that already exists.

Entry points:

• Command palette — the active file must be a type definition in your types folder. It becomes the target.
• File explorer right-click — right-click any type definition file and choose Copy fields from…. That file becomes the target.

Flow:

1. A source-type suggester opens (all types except the target are listed).
2. A dialog shows all fields from the source, with:
   • Fields new to the target — checked by default (will be added).
   • Fields already present in the target — unchecked by default and marked with an overwrite badge. Check them to replace their existing definition.
   • Rename and reorder all included fields before confirming, as in the clone dialog.
3. Click Copy.

Merge behaviour in the target's frontmatter:

• New fields are appended to fields[] and fieldsOrder[].
• Overwritten fields replace their existing definition in place, preserving their position in fieldsOrder[] and their existing Metadata Menu field ID (so MM's index remains consistent).

A notice confirms how many fields were added and how many were overwritten.

Bulk assign object type

Assigns a chosen object type to a scoped set of notes in one operation, with a per-file preview and selection step before anything is written.

Step 1 — Configure

• Object type — the type value that will be written into the type field of every matched file.

Scope

• Directory — vault-relative folder to operate on, with autocomplete. Leave blank for the vault root.
• Recursive — whether to include notes in subdirectories.

Filters

• Include pattern — regex that file paths must match (e.g. ^projects/). Leave blank to include all.
• Exclude pattern — regex that file paths must not match (e.g. ^system/). Leave blank to exclude none.

Post-assign operations (all default to on)

• Clean frontmatter — standard (add missing, preserve everything else) or strict (add missing, remove undefined).
• Sort frontmatter — schema order or alphabetical. When clean is also enabled, sorting is folded into the same pass.
• Reshelve files — move and rename each file to match its new type and title, using the same path/filename mapping logic as the standalone Reshelve command.

All three post-assign operations default to on. Disable any you don't need before clicking Preview.

Step 2 — Preview

A scrollable checklist shows every matched file. Files that already have a type display it as ← current-type so you can spot re-assignments at a glance. All files are checked by default; use Select all / Select none or click individual checkboxes to refine the selection. The summary line updates live.

Clicking Assign commits the operation on the checked files only.

Curate vault (clean / sort / reshelve)

Batch-applies any combination of clean, sort, and reshelve operations across a scoped set of notes. A configuration dialog lets you specify:

Scope

• Directory — vault-relative folder to operate on, with autocomplete. Leave blank for the vault root.
• Recursive — whether to include notes in subdirectories.

Filters

• Include pattern — regex that file paths must match (e.g. ^projects/). Leave blank to include all.
• Exclude pattern — regex that file paths must not match (e.g. ^system/). Leave blank to exclude none.

Operations (enable any combination)

• Clean frontmatter — standard (add missing, preserve everything else) or strict (add missing, remove undefined).
• Sort frontmatter — schema order or alphabetical.
• Reshelve files — move and rename each note to match its type and title.

Clean workspace

Closes all open tabs except one and collapses the file explorer. A quick reset to a clean working state.

Settings reference

Type definitions

Field names

Behaviour

Object catalog defaults

Pre-filled defaults for the Create new workflow object catalog dialog, overridable at creation time.

Path mappings

Controls which vault folder a typed note (object) is filed into, based on its type value. Each mapping is a [regex pattern, LiquidJS template] pair. The first pattern that matches the note's type value wins.

Regex capture groups are available as $1, $2, etc. All frontmatter fields are available as template variables.

Examples

Mappings can be reordered by drag-and-drop. The default maps every type to content/<type>.

Filename mappings

Controls how typed notes (objects) are named, based on their type. Same [regex pattern, LiquidJS template] structure as path mappings. {{ date }} expands to a YYYYMMDDTHHmm timestamp; all frontmatter fields are available as variables.

Examples

The full LiquidJS template syntax is available:

{{ field }}                                        — field value
{{ field | default: "fallback" }}                  — fallback when field is absent
{% if field %}/{{ field }}{% endif %}              — include only when present
{% if flag %}archived/{% else %}live/{% endif %}   — conditional branch
{% if tag %}[{{ tag }}]{% endif %}                 — surround value in brackets

The default maps every type to {{ date }}--{{ title }}.

Acknowledgements

Workflow Objects relies on Metadata Menu by mdelobelle for type schema definitions, and uses LiquidJS for path and filename templating.