skill-name-in-kebab-case

Files (1)

• SKILL.md

SKILL.md (first lines)

# Skill Name

## Overview

Two to four sentences: what this skill does and the outcome it produces. Assume a capable
model — only add what it does not already know.

## When to Activate

Activate when:
- Direct trigger (specific task type or keyword)
- Indirect signal (broader pattern indicating relevance)

**Do not activate** (adjacent skills own this):
- `sibling-skill-a` — owns <nearby job>
- `sibling-skill-b` — owns <nearby job>

## Core Concepts

Only the mental models the agent needs and doesn't already have. Challenge every line for
its token cost.

## Workflow / Detailed Topics

Ordered, executable steps or decision tables. State the freedom level (high / medium / low)
and match specificity to fragility.

1. Step one
2. Step two

For long detail, link out: see [reference](./references/topic.md).

## Practical Guidance

Patterns, anti-patterns, and decision frameworks for applying the skill.

## Examples

**Example:**
```
Input:  <describe input>
Output: <expected output>
```

## Guidelines

1. Verifiable rule one
2. Verifiable rule two

## Gotchas

1. **Title**: what goes wrong and how to prevent it.
2. **Title**: another failure mode and the fix.

## Integration

- `related-skill` — how they hand off (plain text, not a link).

## References

- [Own reference](./references/topic.md)
- Related skill: `skill-name`
- External: spec / docs / paper

From rahulgupta2018/agent-skills, MIT.

More from rahulgupta2018/agent-skills