AI Index
Method and tooling for measuring how AI-generated a piece of prose reads, in Japanese or English. Two layers — a deterministic lint for removable surface tells (scripts/slopscore.py, no API) and a judgment layer of typed probabilistic questions (Jev noul/score) for the tells that survive editing. Includes measured human/AI baselines, a gap-first calibration procedure, and labeled fixtures. Use when building or calibrating an AI-likeness score, or when a draft needs a defensible number rather than a vibe. For writing guidance use natural-writing-ja / natural-writing-en instead.
Install
npx skills add mizchi/skills --skill ai-indexgit clone --depth 1 https://github.com/mizchi/skills.git /tmp/skills
mkdir -p ~/.claude/skills
cp -r /tmp/skills/ai-index ~/.claude/skills/ai-indexPersonal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.
From the SKILL.md
# AI Index A measurement skill: given a text, produce a defensible number for "how AI does this read", and know what that number is worth. The short version of what the measurements below establish: **surface tells are cheap to detect and cheap to remove, so they cannot carry the score.** A text with zero surface tells has had its fingerprints wiped, which is not the same as having something to say. The number that matters comes from the judgment layer. ## When to use - Designing or calibrating an AI-likeness score for drafts, submissions, or CI - Checking a draft before publishing, where "feels AI" needs to become a location and a reason - Auditing a detector someone else built (the gap check in §4 is the audit) Not for: deciding whether a *person* used AI. The measurements here separate unedited model output from human prose. They do not separate a human from a human who edited model output, and §3 shows exactly why. ## 1. The two layers | | Layer A — surface | Layer B — judgment | | --- | --- | --- | | What it finds | word and shape tells (`することができます`, `delve`, uniform sentence length) | whether sentences carry information | | Cost | free, deterministic, no API | one API call per document | | Gap vs unedited AI | **wide** (measured: 0–1 vs 7–8 tells) | +0.77/pt | | Gap vs *edited* AI | **zero** (measured: §3) | **+0.74/pt** via `genericness` (was +0.11/pt) | | Removable by find-and-replace | yes, entirely | no | | Use it as | a lint with a floor | the index |
Details
- Skill name
ai-index- Repository
- mizchi/skills
- Path
ai-index/SKILL.md- Repo updated
- 18 November 2026
FAQ
What does the AI Index skill do?
Method and tooling for measuring how AI-generated a piece of prose reads, in Japanese or English. Two layers — a deterministic lint for removable surface tells (scripts/slopscore.py, no API) and a judgment layer of typed probabilistic questions (Jev noul/score) for the tells that survive editing. Includes measured human/AI baselines, a gap-first calibration procedure, and labeled fixtures. Use when building or calibrating an AI-likeness score, or when a draft needs a defensible number rather than a vibe. For writing guidance use natural-writing-ja / natural-writing-en instead.
How do I install the AI Index skill?
skills CLI (Claude Code, Cursor, Codex and more): npx skills add mizchi/skills --skill ai-index. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.
Is AI Index popular?
It lives in mizchi/skills, which has 423 GitHub stars and 6 forks.
Related skills
Actions CI Tuning
423mizchi/skills
Use when auditing or improving GitHub Actions workflows for a project. Covers cache setup (npm/pnpm/yarn), job parallelism, shard-based test splitting, artifact handling pitfalls, and Playwright-specific patterns. Trigger on: slow CI, cache miss, flaky shard jobs, merge-reports failures, or an explicit 'tune CI' request.
Apm Usage
423mizchi/skills
Reference for APM manifests, installation, targeted updates, deployment paths and lockfiles. Invoke when the user explicitly asks about APM or apm.yml, or an unfamiliar APM command/error needs exact CLI details. Do not automatically invoke for every project that happens to have installed skills.
Ast Grep Practice
423mizchi/skills
Operate ast-grep as a project lint tool. Covers sgconfig.yml, fix/rewrite rules, constraints, transform, testing, and CI. Use when writing rules ast-grep can express but general-purpose linters cannot.
Creating Internal Agents
237medusajs/medusa-agent-skills
Use when building an internal admin-facing AI agent in a Medusa project. These agents are operated by merchants and store operators — not customers. Covers data models, module service, agent runtime (tools, system prompt, streamText), streaming API routes (NDJSON), and admin UI chat extensions. Load for any internal agent type: store operations assistant, product audit, cohort analysis, customer service tooling for support staff, etc. Do NOT use for customer-facing agents (storefront chatbots, buyer-side assistants).
Skill Forge Review
189AgriciDaniel/skill-forge
Audit and validate existing Claude Code skills for quality, triggering accuracy, structure compliance, and best practices. Scores skills on a 0-100 scale and provides prioritized improvement recommendations. Use when user says "review skill", "audit skill", "check skill", "validate skill", or "skill quality".
Empirical Prompt Tuning
423mizchi/skills
Methodology for iteratively improving agent-facing instructions (skills / slash commands / CLAUDE.md / code-gen prompts) via bias-free executor + two-sided evaluation (self-report + instruction-side metrics). Meta-skill, invoke ONLY when the user explicitly asks for an "empirical" eval of a prompt or skill, or for the Iter-0 description / body consistency check. Do NOT auto-invoke after every skill edit; this loop is operator-triggered by name.
Browse the most popular Claude skills, all ai, llms & agents skills, or write your own with the skills generator.