Editscore Ladder Run — Development skill for Claude Code
Run an EditScore-7B pairwise ladder for an image-edit generator (SDEdit / diffusion / LoRA / distilled step-count sweep) on held-out shard pairs.
How to install Editscore Ladder Run
Installs to ~/.claude/skills/v-sekai-fire-dot-claude-editscore-ladder-run/SKILL.md
mkdir -p ~/.claude/skills/v-sekai-fire-dot-claude-editscore-ladder-run && curl -fsSL https://raw.githubusercontent.com/V-Sekai-fire/dot-claude/HEAD/skills/editscore-ladder-run.md -o ~/.claude/skills/v-sekai-fire-dot-claude-editscore-ladder-run/SKILL.md Restart Claude Code, or start a new session, for it to be picked up.
What Editscore Ladder Run does
name: editscore-ladder-run description: Run an EditScore-7B pairwise ladder for an image-edit generator (SDEdit / diffusion / LoRA / distilled step-count sweep) on held-out shard pairs. Use when a new candidate generator needs to be measured against an existing baseline for a shipping-decision (RFD 2186 dressing overlay is the reference case), or when a step-count sweet spot needs to be found before committing distillation compute. Aggregate reports wins / mean / median on the 0-25 overall a
Alternatives in Development
- Skill Creator — Create new skills, modify existing skills, and measure skill performance 94.1k ★
- Breach Check — HIBP k-anonymity check on a password wordlist 4.5k ★
- Fidelity — Measure how faithfully a clone reproduces a site — pixel-diff plus motion-fidelity into one 0-100 score, a let 3.6k ★
Full documentation available on GitHub
View Source RepositoryRelated Skills
Trumps Ten Commandments Skills
Nine Claude agent skills distilled from Jeffrey Sonnenfeld & Steven Tian's Trump's Ten Commandments (2025), us
Warden Select
Measure pending token-warden candidate rules for an agent on the golden suite, evict or activate them, and rec
Warden Power
Zero-token power planner — from the agent's own recorded run-to-run variance, report the minimum detectable sa
Reading Scope
Scope a reading corpus for this writing project — inventory the candidate sources, drop what the project no lo
Fable Eval
Run the fable-mode eval suite (probes → pairwise judge → report). Costs tokens — runs headless claude many tim
Position Ladder
Position Ladder — Staged Entry & Cost-Basis Management
Related Agents
Entity Adjudicator
Decides whether candidate entity/person pairs refer to the same real-world thing, using only the evidence quot
Audit Ecommerce
Audits product page SEO — Product/Offer/AggregateRating schema, pricing consistency, variant handling, breadcr
Name Brand Vetter
Generates and vets candidate product names against the user's own naming taxonomy — a cheap filter on every ge