skills.rest
bastani-inc
bastani-incOFFICIAL
Explore all skills by this author ā†’
WebsiteGitHub

disable-model-invocation

Prevent automatic model selection of manually controlled skills via SKILL.md frontmatter flag.

484|76|Updated Oct 23, 2025
Search Tags:#agent behavior#skill#frontmatter#manual invocation#workflow safety#activation control
One-click install
npx skills add https://github.com/bastani-inc/atomic --skill disable-model-invocation-bastani-inc
Or copy as Structured Prompt for Agentā–¼
Please help me install this Agent Skill.
Skill: disable-model-invocation
Source: https://github.com/bastani-inc/atomic/tree/main/packages/coding-agent/test/fixtures/skills/disable-model-invocation
Command: npx skills add https://github.com/bastani-inc/atomic --skill disable-model-invocation-bastani-inc

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill keeps a skill out of automatic model selection so it can only be triggered when a user explicitly asks for it.

Core Features & Use Cases

  • Manual-only activation: The skill is designed for workflows that should not run unless a human intentionally invokes them.
  • Controlled handling: Useful for sensitive, administrative, or special-case operations that need explicit operator intent.
  • Example: Use it when a tool or procedure should remain hidden from normal model-initiated routing and only run on a direct command.

Quick Start

Use this skill only when you want the workflow to run through the explicit /skill:disable-model-invocation command.

Frequently Asked Questions about disable-model-invocation

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I prevent an agent from automatically triggering a skill?ā–¼

To prevent automatic agent triggering, you can use the disable-model-invocation frontmatter flag to restrict skill activation to explicit user commands only, keeping it hidden from routine model selection.

What is manual invocation control for agent workflows?ā–¼

Manual invocation control is a mechanism that blocks model-initiated skill activation, ensuring sensitive or administrative workflows only execute when an operator explicitly requests them via direct command.

When should I restrict skill activation to explicit user commands?ā–¼

You should restrict skill activation to explicit user commands for sensitive, administrative, or exception-only operations that must remain hidden from normal model-initiated routing and require clear operator intent.

How do I configure the frontmatter to disable model invocation?ā–¼

To configure manual activation, add a frontmatter flag named disable-model-invocation to your SKILL.md file, which limits the workflow to running only through the explicit /skill: command.

Does the disable-model-invocation flag work for sensitive administrative tasks?ā–¼

Yes, the disable-model-invocation flag is designed specifically for sensitive administrative tasks, preventing the model from automatically selecting and running the workflow without explicit operator intent.

Why does my skill keep activating during routine agent selection?ā–¼

Your skill activates during routine selection because it lacks the disable-model-invocation frontmatter flag, which is required to hide workflows from automatic model routing and restrict them to manual commands.

Related Skills

Explore More Software Engineering Skillsā†’
pr-pmpr-pm
121

ci-test-claude-skill

Verify Claude skill installation paths, directory structure, and frontmatter parsing in CI pipelines.

Community
Basic
V1kiV1ki
41

Skill

Invoke the relevant Skill tool as the first action for slash commands.

Community
Basic
ViewableGravyViewableGravy

css_classname-naming

Enforce camelCase CSS class naming in React CSS Modules.

Community
Basic
mbruhlermbruhler
218

using-orchestration

Coordinate multi-agent workflows with sequential, parallel, and conditional execution.

Community
Basic