What is the 'Ask Threshold' used by this skill?

The skill instructs Claude to ask a question if there is a greater than 30% chance of a wrong interpretation or if the cost of an error is significantly higher than the cost of a quick clarification.

How does this skill prevent accidental data loss?

It classifies destructive operations as high-impact ambiguity, requiring explicit user confirmation and documentation of assumptions before any irreversible action is taken.

When will Claude proceed without asking?

Claude will act autonomously for reversible changes, standard industry patterns, or when the intent is clear from the existing project context and the cost of interruption exceeds the cost of a potential correction.

Does this skill support git-based workflows?

Yes, it specifically encourages 'Proceed with Preview' strategies, utilizing git branches, stashes, and dry-runs to ensure all autonomous changes are easily inspectable and reversible.

Decisive Action

Name: Decisive Action
Author: athola

byathola

•

130

•

生産性とワークフロー

Optimizes AI decision-making by balancing autonomous execution with strategic clarifying questions to minimize interaction rounds.

The Decisive Action skill provides a structured framework for Claude to determine exactly when to seek user input versus taking direct action. By implementing a high-impact ambiguity threshold and a reversibility-based decision matrix, it ensures that Claude remains highly efficient without sacrificing safety or correctness. This skill is essential for developers who want to reduce unnecessary 'chattiness' in AI interactions while ensuring that critical decisions—such as destructive operations, security-sensitive changes, or breaking API updates—are always confirmed with the user.

主な機能

01Intelligent Question-to-Action Ratio Balancing

02Industry-Standard Convention Prioritization

03Irreversibility Detection and Confirmation Triggers

04Automated Safety Mechanisms and Preview Protocols

05130 GitHub stars

06Risk-Based Decision Matrix for Destructive Operations

ユースケース

01Speeding up development cycles by allowing autonomous action on reversible git operations.

02Reducing interruptive questions during standard refactoring or formatting tasks.

03Enforcing mandatory confirmation for data deletion or breaking API changes.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add athola/claude-night-market decisive-action

For use in Claude.ai and ChatGPT

主な機能

01Intelligent Question-to-Action Ratio Balancing

02Industry-Standard Convention Prioritization

03Irreversibility Detection and Confirmation Triggers

04Automated Safety Mechanisms and Preview Protocols

05130 GitHub stars

06Risk-Based Decision Matrix for Destructive Operations

ユースケース

01Speeding up development cycles by allowing autonomous action on reversible git operations.

02Reducing interruptive questions during standard refactoring or formatting tasks.

03Enforcing mandatory confirmation for data deletion or breaking API changes.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add athola/claude-night-market decisive-action

For use in Claude.ai and ChatGPT