CertiK Backs Pharos AI Agent Hackathon With Security-Focused Skill Scanner

• The hackathon adheres to a two-stage development approach, with skills coming first and agents coming second, according to the event structure.
• CertiK officially introduced Skill Scanner, a standardized security evaluation tool for AI skill markets, businesses, and developers, in late May.

The largest provider of Web3 security services in the world, CertiK, has joined the Pharos AI Agent Hackathon as an industry sponsor. One of the event’s primary assessment criteria is its recently released Skill Scanner.

The hackathon is a component of Pharos Network’s six-week worldwide AI Agent ecosystem initiative, “Create Like a PRO · Agent Carnival,” which seeks to speed up the creation of safe, production-ready AI Agents and offers a 150,000 PROS prize pool (about $92,500). The hackathon adheres to a two-stage development approach, with skills coming first and agents coming second, according to the event structure.

First, participants will develop standardized skill modules that AI agents may use. After that, these skills will be put together to create fully complete agents that can collaborate, make payments, interact on the chain, and create content. In addition to gaining access to future possibilities within the larger Pharos AI Carnival ecosystem, top-performing projects will compete for a 50,000 PROS prize fund.

According to Pharos, a lot of AI Agent projects nowadays are still primarily demonstration-oriented and lack reusable features and long-term operational procedures. Truly useful Agents should be able to operate for an extended period of time, independently own assets, use tools, and engage in on-chain economic activity rather just acting as one-time displays.

A better developed infrastructure stack for AI agents, comprising standardized skills, security verification procedures, and trusted execution environments, is necessary to realize this ambition.

CertiK officially introduced Skill Scanner, a standardized security evaluation tool for AI skill markets, businesses, and developers, in late May. Through GitHub repositories, URLs, or ZIP packages, users may submit Skills to the platform, which assesses them for risks like as malicious activity, data leaking, unauthorized network access, shell execution, and file system abuse.

With a focus on execution-layer vulnerabilities in financial and automated decision-making contexts, Skill Scanner is intended to evaluate risks that might arise during runtime execution, in contrast to conventional static code analysis tools.

The incorporation of Skill Scanner into the Pharos hackathon’s judging framework is a noteworthy feature that many consider to be one of the first efforts to explicitly include security assessment into AI Agent developer incentive programs.

Hackathon judging has always placed a strong emphasis on technical execution, originality, and user experience. The inclusion of security requirements indicates a change from seeing security as a post-launch issue to a necessity throughout development.

Several AI Agent ecosystems, such as the AI Skill assetization platform FinChip.ai and the AI Skill marketplace Pieverse, have already started to integrate CertiK Skill Scanner. Security solutions like CertiK Skill Scanner are anticipated to become a crucial part of the next generation of AI Agent infrastructure as initiatives like Pharos drive AI Agents toward practical implementation.