Load when creating tutorials, explaining technical concepts, writing documentation, or any content that involves code, data, or technical procedures. Contains guidelines for clear, effective technical communication based on Google's technical writing principles.
This skill inherits all available tools. When active, it can use any tool Claude has access to.
"Every engineer is also a writer." — Google Technical Writing
Technical writing is a learnable skill, not an innate talent. The goal is clear, effective communication that helps readers accomplish tasks.
Avoid → Prefer
utilize → use
leverage → use
in order to → to
prior to → before
subsequent → after
facilitate → help, enable
regarding → about
in the event that → if
# BAD: What the code does (obvious)
x = x + 1 # increment x
# GOOD: Why it's done
x = x + 1 # Account for zero-indexing in display
def calculate_metrics(data: pd.DataFrame, threshold: float = 0.5) -> dict:
"""Calculate accuracy and precision metrics from prediction data.
Args:
data: DataFrame with 'actual' and 'predicted' columns
threshold: Classification threshold for binary predictions
Returns:
Dictionary containing 'accuracy', 'precision', and 'recall' keys
Raises:
ValueError: If required columns are missing
Example:
>>> df = pd.DataFrame({'actual': [1, 0, 1], 'predicted': [0.8, 0.3, 0.6]})
>>> metrics = calculate_metrics(df)
>>> print(metrics['accuracy'])
0.667
"""
# [Tutorial Title]
## What You'll Learn
- Outcome 1
- Outcome 2
- Outcome 3
## Prerequisites
- Requirement 1
- Requirement 2
## Setup
[Environment setup instructions]
## Step 1: [First Action]
[Explanation]
[Code]
[Result]
## Step 2: [Second Action]
[Continue pattern...]
## Complete Example
[Full working code]
## Next Steps
[Where to go from here]
## Resources
[Further reading]
**DataFrame** is like a spreadsheet:
- Rows are individual records
- Columns are fields/variables
- Unlike spreadsheets, operations apply to entire columns at once
Before publishing technical content: