A guide for developing high-quality, token-efficient agent skills with built-in testing and evaluation standards.

Install

mkdir -p .claude/skills/skill-creator-ntluong95 && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/13837" && unzip -o skill.zip -d .claude/skills/skill-creator-ntluong95 && rm skill.zip

Installs to .claude/skills/skill-creator-ntluong95

Activation

This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.

Standards for creating, testing, and optimizing Agent Skills. Use when creating, improving, catching regressions, measuring trigger rates, or writing eval cases for any skill.
175 chars✓ has a “when” trigger
Advanced

Key capabilities

  • Define mandatory frontmatter for skill metadata.
  • Implement a three-level loading system for skill content.
  • Write imperative guidelines using verbs and abbreviations.
  • Adhere to strict size limits for `SKILL.md` and code blocks.
  • Follow strict formatting rules for anti-patterns.
  • Create evaluation cases and design trigger queries for skills.

How it works

The skill provides standards and guidelines for creating, testing, and optimizing agent skills, focusing on token economy, progressive loading, and strict formatting rules. It includes a process for validation and iteration.

Inputs & outputs

You give it
Requirements for a new agent skill or an existing skill for optimization
You get back
A well-structured `SKILL.md` file, evaluation cases, and trigger queries

When to use Skill Creator

  • Create a new agent skill
  • Write evaluation cases for skills
  • Optimize skill documentation
  • Catch skill regressions

About this skill

Agent Skill Creator Standard

Priority: P0 (CRITICAL)

Strict guidelines for High-Density Agent Skills. Maximize info/token ratio.

Core Principles (Token Economy First ⚡)

  • Progressive Loading: Load only essential content initially.
  • Lazy References: Move detailed examples to references/.
  • Imperative Compression: Use verbs, abbreviations, bullet points.
  • Context Limits: Cursor(~100k), Claude(~200k), Windsurf(~32k).

Three-Level Loading System

  1. Metadata: Triggers → AGENTS.md index (Proactive Activation)
  2. SKILL.md: Body < 100 lines → Core guidelines (When triggered)
  3. Resources: references/, scripts/, assets/ → Deep knowledge (On-demand)

Writing Rules

  • Imperative: Start with verbs. "Use BLoC" not "You should use BLoC".
  • Token Economy: Skip articles. Use standard abbreviations. Bullets > paragraphs.
  • Structure:
    1. Mandatory Frontmatter (YAML: name, description, metadata labels & triggers).
    2. Priority: P0 (Critical), P1 (Standard), P2 (Optional).
    3. Guidelines: Imperative Do's.
    4. Anti-Patterns: Strict format Don'ts.
    5. References: Links to lazy-loaded files.

Strict Size Limits

ElementLimitAction if Exceeded
SKILL.md total100 linesExtract to references/
Inline code block10 linesExtract to references/
Anti-pattern item15 wordsCompress to imperative
Tables8 rowsExtract to references/

Strict Formatting Rules

  • Anti-Patterns: **No X**: Do Y[, not Z]. [Context <= 15 words]
    • Example: **No Logic in Builder**: Perform calculations in BLoC, not UI.
  • No Redundancy: Do not repeat frontmatter descriptions.
  • Oversized Skills: If SKILL.md >100 lines, extract step-by-step guides and complex scenarios to references/.
  • Nested Formatting: Avoid **Bold**: \More Bold``.

Test, Measure & Iterate

After writing a skill draft, validate it before shipping:

  1. Write eval cases: Create evals/evals.json with 2–3 realistic prompts (see testing.md).
  2. Design trigger queries: Generate 8–10 should-trigger and 8–10 should-not-trigger queries to measure description accuracy.
  3. Optimize description: Make it "pushy" — list explicit trigger contexts, not just what it does.
  4. Catch regressions: Snapshot the skill before editing; compare before/after outputs for existing test cases.
  5. Iterate: Re-run evals after each change; stop when trigger rate ≥ 80% on held-out queries.

See Testing, Trigger Rate & Regression Guide for eval schema, query design rules, and regression protocol.

Resources & Deep Knowledge

When not to use it

  • When the task does not involve creating, testing, or optimizing agent skills.
  • When ignoring token economy principles for skill content.
  • When not adhering to the three-level loading system.

Limitations

  • The skill enforces strict size limits for `SKILL.md` and inline code blocks.
  • The skill requires specific formatting for anti-patterns.
  • The skill mandates a three-level loading system for content.

How it compares

This skill provides a meta-level framework for skill development, ensuring consistency, efficiency, and testability across all agent skills, unlike ad-hoc skill creation.

Compared to similar skills

Skill Creator side by side with the closest alternatives in the catalog.

SkillInstallsUpdatedSafetyDifficulty
Skill Creator (this skill)05moNo flagsAdvanced
browser-use643moReviewIntermediate
prompt-optimize139moNo flagsAdvanced
ai-cost-optimizer95moCautionIntermediate

Try saying

Example prompts that trigger this skill in your AI assistant.

You might also like

browser-use

browser-use

Automates browser interactions for web testing, form filling, screenshots, and data extraction. Use when the user needs to navigate websites, interact with web pages, fill forms, take screenshots, or extract information from web pages.

64174

prompt-optimize

YYH211

Expert prompt engineering skill that transforms Claude into "Alpha-Prompt" - a master prompt engineer who collaboratively crafts high-quality prompts through flexible dialogue. Activates when user asks to "optimize prompt", "improve system instruction", "enhance AI instruction", or mentions prompt engineering tasks.

13113

ai-cost-optimizer

ScientiaCapital

Save 40-70% on AI costs with intelligent multi-LLM routing. Automatically selects the optimal model based on task complexity across 40+ models from 8 providers.

990

tmux-terminal

mikeyobrien

Interactive terminal control via tmux for TUI apps, prompts, and long-running CLI workflows.

319

self-improving-agent

alirezarezvani

Curate Claude Code's auto-memory into durable project knowledge. Analyze MEMORY.md for patterns, promote proven learnings to CLAUDE.md and .claude/rules/, extract recurring solutions into reusable skills. Use when: (1) reviewing what Claude has learned about your project, (2) graduating a pattern from notes to enforced rules, (3) turning a debugging solution into a skill, (4) checking memory health and capacity.

128

ccs-delegation

kaitranntt

Auto-activate CCS CLI delegation for deterministic tasks. Parses user input, auto-selects optimal profile (glm/kimi/custom) from ~/.ccs/config.json, enhances prompts with context, executes via `ccs {profile} -p "task"` or `ccs {profile}:continue`, and reports results. Triggers on "use ccs [task]" patterns, typo/test/refactor keywords. Excludes complex architecture, security-critical code, performance optimization, breaking changes.

210

Search skills

Search the agent skills registry