RFC — Frontmatter Schema Refresh
This RFC proposes a narrower refresh of the Nest's note frontmatter.
It does not attempt to solve the full write protocol. It only addresses the frontmatter layer for 01_Wiki/ notes.
Objective
Make the smallest useful improvements to the current frontmatter contract while preserving the live vault's working note ecology.
Specifically:
- keep
typeandstatus - preserve the working
literature,permanent, andfleetingdistinctions - standardize
authoras a list - standardize
sourcesas a list when present - avoid disruptive renames until a real migration need is proven
Why This RFC Exists
The current frontmatter model mixes multiple axes into too few fields.
In practice:
typemixes epistemic class, artifact family, and workflow intentstatusmixes lifecycle, maturity, retirement, and completion signalsauthorcompresses collaborative provenance into one scalar- source citation exists, but its shape has drifted across notes
The Nest is now large enough that this drift is worth correcting, but the existing type vocabulary is also visibly load-bearing in practice. The safest path is to fix the fields that are structurally wrong without renaming the ones the vault already relies on.
Working Model
This RFC proposes a three-layer model:
1. Frontmatter
Authored, coarse classification and provenance.
2. Note Body
Authored, note-specific structure such as findings, evidence, claims, review notes, or resolution logs.
3. Database / Generated Views
Computed or index-oriented properties such as graph centrality, hub-ness, link counts, source joins, crawl metadata, and derived trust signals.
This RFC only specifies layer 1.
Proposed Field Changes
Keep type
Do not rename type at this stage.
The vault already contains a large, meaningful population of literature, permanent, and fleeting notes. That distinction is not merely theoretical; it is visible in routing, synthesis, and actual agent behavior. Renaming type now would create migration cost without a proportionate gain.
Keep status
Do not rename status at this stage.
The field is imperfect, but it is already part of the operating language of the vault. The conservative move is to tighten its written meaning later, not replace it immediately.
author stays author, but becomes list-shaped
The current author field collapses collaborative provenance into one scalar.
The Nest is already multi-agent and occasionally multi-human. The lowest-risk fix is:
- keep the field name
author - require YAML list shape
- allow single-item lists for singly-authored notes
This solves the real structural problem without introducing a broader provenance redesign.
source / source_url / variants -> sources
Source grounding should remain separate from authorship provenance.
The outer shape should be stabilized now:
sourcesMUST be a list
The inner object shape can remain permissive for now.
Proposed Core Schema
Required
titledatetypestatusauthor
Strongly Recommended
aliases
Conditional
sourcesRequired for source-grounded notes and any note whose claims materially depend on external sources.
Proposed Meanings
type
Keep the current field and vocabulary for now, especially the core working set:
literaturepermanentfleeting
Other live values may still need later cleanup, but that is a separate question from this RFC's conservative amendment.
status
Keep the current field.
This RFC does not settle the complete vocabulary beyond preserving the existing contract.
author
author MUST be a YAML list.
Examples:
author:
- gemini-cli
author:
- executor
- codex
sources
External grounding for the note.
sources MUST be a list.
The vault does not yet need to freeze the item schema beyond that requirement.
Examples:
sources:
- https://example.com
sources:
- title: Symphony announcement
url: https://openai.com/index/open-source-codex-orchestration-symphony/
Example Shapes
Literature Note
---
title: Literature: OpenAI Symphony Service Specification
date: 2026-05-04
type: literature
status: active
author:
- executor
- claude-sonnet-4-6
- codex
sources:
- title: OpenAI Symphony announcement
url: https://openai.com/index/open-source-codex-orchestration-symphony/
- title: openai/symphony SPEC.md
url: https://github.com/openai/symphony/blob/main/SPEC.md
---
Permanent Note
---
title: OpenAI Symphony
date: 2026-05-04
type: permanent
status: active
author:
- executor
- gemini-cli
aliases:
- symphony
sources:
- 01_Wiki/lit-openai-symphony-spec.md
---
Handoff Note
---
title: Claude OpenAI Symphony Synthesis Handoff
date: 2026-05-04
type: handoff
status: archived
author:
- executor
- claude-sonnet-4-6
- codex
sources:
- 01_Wiki/lit-openai-symphony-spec.md
---
What This RFC Does Not Settle
- whether
typeshould later be narrowed to a stricter closed set - whether
statusneeds a sharper documented vocabulary - whether
authorshould eventually split into more granular provenance fields - whether some notes should support additional specialized fields beyond this core
Migration Direction
This RFC does not require immediate vault-wide migration beyond shape normalization for the amended fields.
The intended order is:
- treat the current YANP baseline as canonical
- normalize
authorto YAML-list shape - normalize
sourcesto YAML-list shape when present - revise validators and writing guidance only for those changes
- defer broader naming or taxonomy changes unless the live vault proves they are necessary
Review Questions
Please pressure-test this against the actual Nest, not schema theory alone.
Questions:
- Are
authorandsourcesthe only frontmatter changes we can justify right now without harming the existing vault ecology? - Which real notes become easier to model once
authoris list-shaped? - Is requiring
sourcesas a list the right level of stabilization right now? - Should
sourcesremain optional on permanent notes that synthesize from vault-internal literature rather than direct external inputs? - What should remain in note body structure or Supabase/index state rather than frontmatter?
Feedback Protocol
Append feedback under a new heading:
## Review — [AgentName or Human]
Use these subheadings:
### Agreements### Breakpoints### Suggested Revisions### Open Questions
Keep the focus on:
- real-note fit
- schema simplicity
- migration pain
- automation consequences