Agent Skills: Using Skills

Guidelines for using skills effectively - load relevant skills before complex tasks, not every message

UncategorizedID: Cygnusfear/claude-stuff/using-superpowers

Install this agent skill to your local

pnpm dlx add-skill https://github.com/Cygnusfear/claude-stuff/tree/HEAD/skills/using-superpowers

Skill Files

Browse the full folder contents for using-superpowers.

Download Skill

Loading file tree…

skills/using-superpowers/SKILL.md

Skill Metadata

Name
using-superpowers
Description
Guidelines for using skills effectively - load relevant skills before complex tasks, not every message

Using Skills

Skills are specialized workflows that improve quality for specific task types. Use them when they apply - don't skip them by rationalizing, but also don't invoke them ritualistically.

When to Load a Skill

DO load skills when:

  • Starting a complex or multi-step task
  • The task clearly matches a skill's description (code review, debugging, planning, etc.)
  • You're unsure how to approach something and a skill might help
  • The user explicitly requests a skill-based workflow

DON'T load skills when:

  • Answering simple questions or having conversation
  • Doing trivial file reads or small edits
  • The task is straightforward and no skill adds value
  • You already know the skill's content from this session

How to Use Skills

When a skill applies:

  1. Load it with the Skill tool
  2. Announce briefly: "Using [skill] for [purpose]"
  3. If it has a checklist, create todos
  4. Follow the skill's workflow

Avoiding Common Mistakes

Under-using skills (rationalizing):

| Thought | Consider | |---------|----------| | "I'll just do this quickly" | Would a skill improve quality? | | "This doesn't need a formal process" | Is there a skill that applies? | | "I remember how this works" | Skills evolve - reload if unsure | | "Let me explore first" | Some skills guide exploration |

Over-using skills (ritual compliance):

| Thought | Reality | |---------|---------| | "I must invoke skills on every message" | Only when they add value | | "Let me check for mandatory skills" | Skills aren't rituals | | "Confirming skill protocol compliance" | Just do the work |

Skill Priority

When multiple skills could apply:

  1. Process skills first (brainstorming, debugging) - these determine HOW to approach
  2. Implementation skills second - these guide execution

Examples:

  • "Let's build X" → brainstorming first, then implementation
  • "Fix this bug" → debugging first, then domain-specific skills

Skill Types

Rigid (TDD, debugging): Follow the process exactly.

Flexible (patterns): Adapt principles to context.

The skill itself indicates which type it is.

Remember

  • Skills improve quality for complex tasks
  • Simple tasks don't need skill overhead
  • If you're announcing "checking mandatory protocols" you've gone too far
  • Just use skills naturally when they help