action-validator

Name: action-validator
Availability: InStock
Author: hiddink-ai

Validate agent tool calls against declared capabilities and file access scopes.

1|Updated May 21, 2026

Search Tags:#validation #compliance #safety #guardrails #tool-use #agent-orchestration

Ready to run

npx skills add https://github.com/hiddink-ai/hiddink-harness --skill action-validator-hiddink-ai

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: action-validator
Source: https://github.com/hiddink-ai/hiddink-harness/tree/main/templates/skills/action-validator
Command: npx skills add https://github.com/hiddink-ai/hiddink-harness --skill action-validator-hiddink-ai

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill addresses the risk of AI agents performing unauthorized or out-of-scope actions by providing an advisory validation layer that monitors tool calls against defined task contracts and file access permissions.

Core Features & Use Cases

Boundary Enforcement: Validates tool usage against an agent's specific tool list and domain scope.
Policy Caching: Captures and reuses validated decision paths for repetitive workflows to improve efficiency and consistency.
Advisory Warnings: Provides non-blocking, real-time feedback when an agent attempts an action that deviates from its declared capabilities or safety rules.

Quick Start

Enable the action validator to monitor the current agent session and provide warnings for any tool calls that fall outside the defined task scope.

Frequently Asked Questions about action-validator

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

FAQPage Schema

How do I prevent AI agents from performing unauthorized tool calls during execution?▼

To prevent unauthorized tool calls, you can use an advisory validation layer that monitors agent actions against declared capabilities and file access scopes, providing real-time warnings for boundary violations during execution.

What is agent tool-call scope validation and how does it work?▼

Agent tool-call scope validation is an advisory mechanism that checks tool usage against defined task contracts and domain boundaries. It operates within the agent harness using policy caching to ensure operational legality.

How do I enforce file access boundaries and task contracts for autonomous agents?▼

You enforce file access boundaries by integrating a validation guard with the agent's tool-use hooks and policy cache. This setup validates execution attempts against task contracts and provides non-blocking advisory feedback.

Do I need agent orchestration hooks to enable real-time safety guardrails?▼

Yes, enabling real-time safety guardrails requires integration with the agent's tool-use hooks and policy cache. This connection allows the validation layer to intercept actions and enforce boundary constraints.

Can safety guardrails block an agent's tool execution if it violates policy?▼

No, these safety guardrails operate as a non-blocking advisory layer. They detect scope violations and provide actionable feedback during execution, but they do not hard-block the agent's tool calls.

What is the best way to monitor repetitive agent workflows for compliance?▼

The best way to monitor repetitive workflows for compliance is using policy caching, which captures and reuses validated decision paths to improve efficiency and maintain consistent boundary enforcement.