prempti-falco-rules

Author custom Falco rules to monitor AI coding agent tool calls.

193|22|Updated Mar 18, 2026
One-click install
npx skills add https://github.com/falcosecurity/prempti --skill prempti-falco-rules
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: prempti-falco-rules
Source: https://github.com/falcosecurity/prempti/tree/main/skills/prempti-falco-rules
Command: npx skills add https://github.com/falcosecurity/prempti --skill prempti-falco-rules

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill addresses the security risks posed by AI coding agents by providing a policy and visibility layer that intercepts tool calls, preventing unauthorized file access, destructive shell commands, and credential exfiltration.

Core Features & Use Cases

  • Real-time Guardrails: Enforce allow, deny, or ask-for-confirmation verdicts on every tool call made by agents like Claude Code.
  • Custom Policy Authoring: Create and validate Falco rules tailored to your specific environment and security requirements.
  • Use Case: Prevent an AI agent from accidentally running destructive commands like rm -rf or accessing sensitive configuration files like .env or .ssh/id_rsa by defining custom deny rules.

Quick Start

Use the prempti-falco-rules skill to draft a new rule that blocks the agent from executing any shell command containing the word sudo.

Frequently Asked Questions about prempti-falco-rules

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

FAQPage Schema
How do I enforce security guardrails on AI coding agents?

You can enforce security guardrails on AI coding agents by authoring custom Falco rules that intercept tool calls and apply allow, deny, or ask-for-confirmation verdicts to prevent unauthorized actions.

How do I write a Falco rule to block an AI agent from running destructive shell commands?

To write a Falco rule blocking destructive shell commands, define a custom deny rule adhering to the coding_agent event source schema, targeting specific command patterns like rm -rf to stop execution.

Can I prevent AI coding agents from accessing sensitive files like .env?

Yes, you can prevent AI coding agents from accessing sensitive files like .env or .ssh/id_rsa by creating custom Falco rules that deny file access attempts and log the activity for auditing.

Does this Skill require a specific CLI tool to install and manage Falco rules?

Yes, this Skill requires manual installation and management of the Falco rules via the premptictl CLI tool to enforce policy and monitor tool-call activity for your agentic workflows.

What is the coding_agent event source schema used for in Falco rules?

The coding_agent event source schema defines the specific data structure your custom Falco rules must adhere to when intercepting and evaluating tool calls made by AI coding agents.

How does Falco handle credential exfiltration attempts by AI agents?

Falco handles credential exfiltration by intercepting the agent's tool calls, evaluating them against your custom deny rules, and blocking the action while generating an audit log for threat mitigation.