dp-meta-layer-draft

name

dp-meta-layer-draft

description

Drafts Desirable Property text and inscription-ready variants from app.themetalayer.org DP sources, applying the full meta-layer lens. Saves outputs to the agent-lab Draft Editor via save_draft when available. Produces article-first or video-first companion assets and CTAs for comments and supporting/opposing artifacts. Use when working on DP drafts, on-chain inscription text, gov hub copy, meta-layer properties, Draft Editor, or "draft DPn / desirable property / inscription."

Desirable Property — meta-layer draft & inscription

Goal

Turn existing DP material (typically https://app.themetalayer.org/?dp=N plus optional pasted notes) into:

1. A full property draft suitable for the gov hub (clear, reviewable).
2. An inscription draft — a tighter, commitment-oriented variant for on-chain or equivalent publication (user confirms chain/tool constraints).
3. Comms runway: article and/or video-oriented copy that invites feedback and linked artifacts.

When to use

• User names a DP number or link (dp=1 … dp=21).
• User asks for inscription text, gov hub draft, or meta-layer-aligned rewrite of a property.
• User points at examples/dp1-reference.md as the style anchor (DP1 reference is the default depth model).

Inputs

1. DP source URL — fetch app.themetalayer.org/?dp=N (and themetalayer.org/desirable-properties for category context if useful).
2. Optional: user paste, meeting notes, or challenge context (e.g. external critique).
3. Style anchor: examples/dp1-reference.md is populated (full DP1 ML-Draft). Match its structure and depth for DP2–DP21 unless the user asks for a shorter pass.

Meta-layer lens

Read and apply meta-layer-lens.md as a full pass: identity/agency, sovereignty, interoperability, AI governance, trust/provenance/sustainability, community feedback. Surface tradeoffs and open questions explicitly.

Comms order (two modes)

Mode	Order	Use when
Runway (default)	Article → video	Building initial library; article grounds the draft; video amplifies.
Video-first	Video → article	Higher video volume; short-form leads; article becomes expanded record.

State which mode applies in the deliverable header. User can override per request.

Deliverables

Produce all of the following unless the user restricts scope:

A. Full property draft (gov hub)

• Title line: DPn — [short name]
• Summary (2–4 sentences)
• Statement of the property (precise, implementable where possible)
• Rationale — why this matters for a human-first meta-layer
• Implications — builders, users, governance
• Open questions / iteration — what feedback would change the draft

B. Inscription draft

• Standalone text that can ship as a single inscription (no reliance on unstated context).
• Shorter than the full draft; commitment tone (what is being claimed or recorded).
• If character/word limits are unknown, produce one primary + one ultra-tight variant and note "trim to chain limits."
• Mark status: e.g. draft for review — final wording is human-approved before inscription.

C. Article companion (runway)

• Title options (2)
• Lead (2–3 paragraphs) usable in Substack or site
• CTA block (must include):
  • Invite comments on the draft
  • Invite links to supporting or opposing artifacts (papers, posts, code, governance decisions) with short guidance on what makes a useful link

D. Video companion (runway or video-first)

• 60–90s beats: hook → what this DP is → why it matters → what feedback we need
• On-screen title suggestion
• Spoken CTA: comment + link an artifact (same intent as article)

E. Cross-DP consistency

• Name adjacent DPs that this draft touches or must stay aligned with (same page/category).

Draft Editor (agent-lab)

Yes — these drafts belong in the Draft Editor the same way other agent output does. The DeerFlow save_draft tool writes to data/draft_store/; the UI (typically http://localhost:8081) lists them for review and publish.

When save_draft is available

Use one save_draft call per logical artifact (do not cram unrelated variants into one draft). Typical mapping:

Artifact	platform	destination	title (short label)
Full gov-hub draft (A)	substack	Substack · metaweb	DPn — full draft (gov hub)
Inscription draft (B)	substack or linkedin	Same family as above, or org LinkedIn if inscription is staged as a post	DPn — inscription draft
Article companion (C)	substack	Substack · metaweb	DPn — article / Substack
Video beats (D)	linkedin	LinkedIn · Meta-Layer Initiative or personal — pick one	DPn — video beats

Use metadata_json so the editor and future you can filter. Merge with any other metadata keys the tool expects; at minimum include:

{
  "title": "<same as save_draft title param>",
  "draft_type": "desirable_property",
  "dp_number": <integer>,
  "variant": "full_gov_hub | inscription | article | video_beats"
}

Optional: "source_url": "https://app.themetalayer.org/?dp=N".

Publish profiles: Use the exact destination strings from the user’s data/shiftshapr_context.json → publish_profiles (or the agent context block) — same as meta-layer-article / event promo flows.

After saving, tell the user: Review in Draft Editor (local URL if known).

When save_draft is not available

(e.g. Cursor without DeerFlow / AGENT_LAB_ROOT unset)

1. Preferred: Run the same workflow through Shiftshapr / DeerFlow so save_draft executes in agent-lab context.
2. Fallback: From the agent-lab repo root, with PYTHONPATH including the project root, the agent may call apps.draft_editor.store.create_draft in a short script — only if the user wants files landed without DeerFlow; otherwise output Markdown in chat for manual paste.

Workflow

1. Fetch DP source; extract claims, lists, and definitions.
2. Apply meta-layer-lens.md; note gaps vs. lens.
3. If dp1-reference.md is populated, align structure to it.
4. Draft A → B → C → D in that order (inscription after full draft so it stays faithful).
5. Add CTA blocks to C and D as specified.
6. If save_draft is available, persist A–D (and any promo variants) to the Draft Editor per the table above; include metadata_json for desirable_property drafts.

Anti-patterns

• Do not invent Meta-Layer Initiative commitments the user did not state.
• Do not skip the artifact CTA — it is part of the program.
• Do not merge inscription and long draft into one blob without a labeled inscription section.