Gating rules, scoring behavior, and system preferences.
GATING RULES
Define when content promotes, demotes, or gets killed. Applied automatically.
Loading rules…
KEYBOARD SHORTCUTS
Open command palette⌘K
Dashboard⌥1
Inbox⌥2
Studio⌥3
Themes⌥4
Gating⌥5
Publish⌥6
Submit (in Capture form)⌘↵
Indent (in Editor)Tab
Send (in AI chat)↵
OBSIDIAN EXPORT
Export all ideas and documents as Markdown files in Obsidian-compatible folder structure. Drop into your vault under a content-os/ folder.
SYSTEM INFO
appwriting-mgmt
convention2026.05
supabaseinfluence-mxf
aigemini-2.0-flash
frameworknext.js 16.2
aestheticTrack C · Rorra
DRAFT SCORING LOGIC
Used to gate Private → Draft. A draft needs 70 / 100 to advance. Scored by Claude Opus on 4 dimensions (25pts each). Edit this section to revise the criteria passed to the model.
ADVANCEMENT THRESHOLD
Score ≥ 70 to move from Private to Draft
70
SCORE LEVELS
0 – 39Raw MaterialNo clear claim, vague, generic.
40 – 54Taking ShapeA direction is visible but argument has gaps.
55 – 69Almost ThereClose but one dimension is pulling it down.
70 – 84Solid Draft✓ Gate passed. Ready to advance to Draft.
SCORING DIMENSIONS — 25 pts each · model: claude-opus-4-5
Claim Clarity0 – 25 pts
Is there one specific, arguable claim the reader walks away believing? Not a topic — a claim. A piece 'about productivity' scores low. Penalize pieces that survey without landing on a position.
Specificity0 – 25 pts
Are there concrete numbers, named examples, or specific observations? 'Many companies have seen significant impact' scores low. '3x value' without context still scores low — specificity requires the reader to be able to verify or picture the claim.
Argument Coherence0 – 25 pts
Does each section build on the last? Cause-effect chains must hold. No logical leaps, no unsupported transitions, no conclusions that don't follow from the evidence presented. Structural integrity, not prose quality.
Original Voice0 – 25 pts
Does this sound like a specific person with a specific POV? Penalize: corporate hedging, passive constructions, AI-summary tone (balanced, comprehensive, nuanced), and interchangeable sentences that could have been written by anyone.