Skip to content
Back to skills

Using Vibekit

ASecurity

Use when starting any conversation: establishes the auto-trigger discipline so guardrail skills fire instead of being silently skipped.

  • 20 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added October 8, 2026
ai-agentsgo

Security analysis

A100/100

Scanned October 8, 2026

npx -y skills add rizukirr/vibekit --skill using-vibekit --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Using Vibekit?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Using Vibekit
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/rizukirr-using-vibekit/badge)](https://www.skillsdirectory.com/skills/rizukirr-using-vibekit)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: using-vibekit
description: "Use when starting any conversation: establishes the auto-trigger discipline so guardrail skills fire instead of being silently skipped."
trigger: Session start
gate: none
---

# using-vibekit

If there is even a 1% chance a vibekit skill applies to what you are about to do, invoke it.

This is not negotiable. "The task is too small", "I already know the answer" and "it would be faster to just do it" are the rationalisations this plugin exists to stop. A guardrail you talked yourself out of is a guardrail that was never there.

If a skill turns out to be wrong for the situation, you do not have to follow it, but you do have to check.

## If you are a subagent

If you were dispatched to execute a specific task, skip this and follow your brief. The orchestration discipline belongs to the session that dispatched you.

## Instruction priority

1. **The user's explicit instructions**: highest. If they say "skip the design step", skip it.
2. **vibekit skills**: these override default behaviour where they conflict.
3. **The default system prompt**: lowest.

## Finding the right skill

Every skill declares its own trigger, and the auto-trigger table in `CLAUDE.md` is generated from those declarations, so it is never out of date. Read the table, not a copy of it.

A skill whose row says `hard` is a gate. Respect it regardless of how simple the task looks. Simple tasks are where unexamined assumptions cost the most.

## Always on

Some skills are modifiers rather than steps: they say so in their own descriptions, which end "Stays on after." The trigger table lists them like any other skill. Nothing here enumerates them, because a count kept by hand goes stale the moment one is added.

Invoke each **once**, at the first moment its trigger applies, then keep applying it for the rest of the session without invoking it again. A modifier you never invoke is a modifier whose content you never read: the description line alone is not the skill.

## How to invoke

Use the `Skill` tool. The skill's content loads and you follow it directly. Never read a skill file as a substitute for invoking it: reading gives you the text without the commitment.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.