Securing the Agent Skills Registry: How Snyk and Tessl Are Setting the Standard

2026-03-17 · Source: Blog RSS Feed | Snyk · Field: Technology & Digital — Cybersecurity & Data Privacy, Artificial Intelligence & Machine Learning, Software Development & Engineering · Depth: Advanced, medium

Summary

Snyk and Tessl have partnered to integrate security scanning into the Tessl Registry, addressing the nascent and vulnerable ecosystem of AI agent skills. Unlike traditional software packages, agent skills are natural language instructions that guide autonomous AI agents, creating a distinct attack surface. Snyk's ToxicSkills research, which scanned 3,984 skills from the ClawHub marketplace, revealed that 36% contained prompt-injection techniques, with all confirmed malicious skills combining prompt injection and malicious code payloads. The new integration displays Snyk security scores directly on Tessl Registry skill pages and in search results, and the Tessl CLI will warn developers about known issues. This system, powered by Snyk's agent security technology, analyzes behavioral intent for threats like prompt injection, malware, and toxic flows, aiming to establish trust signals early in the agent skills supply chain.

Key takeaway

For AI Engineers or Software Engineers evaluating agent skills, you must recognize that these components present unique security risks beyond traditional code. Your decision to install a skill should now incorporate Snyk's security scores visible in the Tessl Registry and CLI. Proactively run Snyk's "agent-scan" locally against your configurations to identify prompt injection or toxic flow patterns, mitigating potential agent goal hijack and data exfiltration risks.

Key insights

AI agent skills introduce unique security risks, requiring specialized scanning beyond traditional code vulnerability detection.

Principles

• Agent skills' natural language instructions create a new attack surface.
• Autonomous agents shift the trust model, increasing risk.
• Security must be embedded at discovery and installation.

Method

Snyk's agent security technology uses calibrated models and deterministic rules to analyze behavioral intent, checking for prompt injection, malware, credential mishandling, and toxic flow patterns.

In practice

• Browse Tessl Registry for Snyk security scores.
• Run Snyk's "agent-scan" CLI locally.
• Review ToxicSkills research for threat taxonomy.

Topics

• AI Agent Security
• Agent Skills Registry
• Snyk
• Tessl
• Prompt Injection
• Supply Chain Security

Code references

• snyk/agent-scan
• mcollina/skills
• better-auth/skills
• Jeffallan/claude-skills
• kepano/obsidian-skills

Best for: CTO, VP of Engineering/Data, Director of AI/ML, AI Security Engineer, AI Engineer, Software Engineer

Related on AIssential

• vercel-labs / skills — The "skills" CLI centralizes management of reusable agent instruction sets across diverse coding agents.
• How Relay Network Adopted AI Coding Securely and Built the Foundation for Agentic Development — Embedding security "secure at inception" into AI-assisted development prevents vulnerabilities and accelerates remediation.
• The AI Agent Security Surface: What Gets Exposed When You Add Tools and Memory — AI agents introduce four distinct attack surfaces requiring a new security paradigm beyond traditional LLM prompt defenses.
• Practical Guide to Evaluating and Testing Agent Skills — Systematic evaluation of agent skills is crucial for reliability, performance, and cost-efficiency.
• Fragments: April 21 — AI's rapid advancement necessitates revisiting foundational software principles and developing robust security for "permission hungry" agents.
• Anatomy of Agent SKILLS — Agent "skills" prevent context bloat by progressively disclosing information, improving efficiency and modularity.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Blog RSS Feed | Snyk.