skills.rest
Ghenghis
GhenghisCOMMUNITY
Explore all skills by this author →
GitHub

yolo-permissions

Classify tool-call risk before execution using quick rules and an AI classifier.

Updated Apr 3, 2026
Search Tags:#bash#safety#ai-security#risk-classification#yolo#tool-call
One-click install
npx skills add https://github.com/Ghenghis/openclaw-claude-code-integration --skill yolo-permissions-ghenghis
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: yolo-permissions
Source: https://github.com/Ghenghis/openclaw-claude-code-integration/tree/main/skills/yolo-permissions
Command: npx skills add https://github.com/Ghenghis/openclaw-claude-code-integration --skill yolo-permissions-ghenghis

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Tool invocations in AI agents can trigger unsafe actions if not properly vetted. YOLO permissions acts as a safety layer that classifies the risk of every tool call before it runs, preventing dangerous operations and guiding the agent with clear remediation steps.

Core Features & Use Cases

  • Risk-aware gating: evaluates tool calls (bash commands, file edits, external communications) and blocks or flags high-risk actions.
  • Multi-layer decision: combines quick static rules with an AI-assisted classifier for robust coverage.
  • Real-world use: protects automated workflows in development environments, CI pipelines, and automated agents from executing destructive or credential-leaking operations.

Quick Start

Provide a tool call to be evaluated and receive a risk level, rationale, and recommended action.

Frequently Asked Questions about yolo-permissions

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

FAQPage Schema
How do I classify tool-call risk before execution to prevent unsafe actions?▼

Classifying tool-call risk before execution prevents unsafe actions by evaluating bash commands, file edits, and external communications through a two-layer model combining static rules and an AI-backed classifier to return structured risk levels and remediation steps.

What is a two-layer risk model for AI agent safety?▼

A two-layer risk model for AI agent safety combines quick static rule checks with an AI-assisted classifier to robustly evaluate automated tool invocations, returning structured risk, reasoning, and recommended actions to gatekeep dangerous operations.

How do I gatekeep dangerous bash commands in automated workflows?▼

Gatekeeping dangerous bash commands in automated workflows requires a safety layer that evaluates tool calls before they run, blocking destructive operations or credential-leaking actions while guiding the agent with clear remediation steps.

Can I use risk classification to protect CI pipelines from destructive operations?▼

Risk classification can protect CI pipelines from destructive operations by evaluating automated tool invocations across development environments, blocking high-risk bash commands, file edits, and external communications before execution.

Does automated tool invocation risk gating work without external dependencies?▼

Automated tool invocation risk gating works without external dependencies by implementing a self-contained two-layer model that applies quick-rule checks and an AI-backed classifier to gatekeep potentially dangerous actions.

Related Skills

Explore More Software Engineering Skills→
itsadijmbtitsadijmbt
1

tool-selection

Evaluate Databricks agent traces for MCP tool selection over Bash workarounds.

Community
Intermediate
render-ossrender-oss

render-debug

Diagnose Render deployment failures by analyzing build, deploy, and runtime logs.

Official
Advanced
thistleknotthistleknot

agent-governance

Enforce deny-by-default policy rules on agent tool calls with audit logging.

Community
Advanced
zaeem-rafiqzaeem-rafiq

git-guardrails

Block destructive git operations with pre-commit and pre-push hooks.

Community
Advanced