Golden template

{Skill Display Name}

Quick Start

{2-3 sentences: the core workflow in the simplest possible terms.}

Workflow

Step 1: Gather Context

{Concrete instruction with specific action — imperative voice.}

• For {scenario A}, read ${CLAUDE_SKILL_DIR}/references/detailed-guide.md
• For {scenario B}, read ${CLAUDE_SKILL_DIR}/references/examples.md

Step 2: Execute

{Concrete instruction — what to do, not what to think about.}

• Run: python ${CLAUDE_SKILL_DIR}/scripts/main_helper.py --input "$ARGUMENTS"

Step 3: Validate

{Concrete instruction — verify the output meets quality bar.}

• Run: bash ${CLAUDE_SKILL_DIR}/scripts/validate.sh
• If validation fails, fix issues and re-validate before presenting to user.

Examples

Example 1: {Typical use case} Input: "{realistic user prompt, conversational tone}" Output: {expected result with specific format}

Example 2: {Edge case} Input: "{edge case prompt that should still trigger}" Output: {expected result}

Example 3: {Near-miss — should trigger but might not} Input: "{prompt using synonym/adjacent concept}" Output: {correct handling}

Gotchas

• {Known failure point #1 — what goes wrong and how to handle it}
• {Known failure point #2 — specific, actionable}
• {Common user mistake and correct response}

Out of Scope

This skill does NOT:

• {Explicit exclusion} (use {other-skill} instead)
• {Explicit exclusion}
• {Explicit exclusion}