Blog | Jiffy Labs

OWASP published the Agentic Skills Top 10 in August 2026. What each of the ten risks says, the evidence behind them, and where security teams should start.

Jiffy Labs Research

•

August 18, 2026

•

min read

owasp-agentic-skills-top-10

Engineering

Where AI Artifact Scanning Belongs in Your CI/CD Pipeline

Static, dynamic, and reasoning-based scanning carry different costs and latency. Here's where each one actually belongs in your delivery pipeline.

ai-artifact-scanning-cicd-pipeline

Industry Analysis

From Endpoint to Governance: Defense in Depth for the AI Supply Chain

Point solutions catch point problems. Defense in depth for the AI supply chain means inspecting the artifact itself at every point it enters your environment.

defense-in-depth-ai-supply-chain

Threat Research

Anatomy of a Real Attack: Dissecting a Data-Exfiltration Agent Skill

A walkthrough of a real malicious agent skill from our research dataset, from its innocent-looking description to the socket connection that gave it away.

anatomy-of-a-malicious-agent-skill

Product

Why Every Verdict Now Ships With Its Evidence

A score with nothing behind it gets argued with once and ignored after that. Here's what changed when we made every verdict carry the evidence that produced it.

Jiffy Labs Engineering

verdicts-ship-with-evidence

Engineering

One Verdict Isn't Enough: The Case for Hybrid Agent Skill Scanning

Static rules, sandboxed execution, and reasoning over evidence each catch something the other two miss. None of them is sufficient alone, and our own benchmark shows exactly why.

hybrid-approach-agent-skill-scanning

Policy & Compliance

Where AI Artifacts Fit in the Compliance Framework You Already Have

You probably don't need a new control framework for agent skills. You need to notice that several controls you already run were written as though this category didn't exist.

ai-artifacts-existing-compliance-framework

Engineering

Inside the Sandbox: What Dynamic Analysis Reveals About Agent Skills

Static review reads what a skill says it will do. A sandbox shows what it actually does the moment it runs.

dynamic-analysis-agent-skills-sandbox

Engineering

Why Static Analysis Alone Can't Catch Malicious Agent Skills

Signature and pattern matching are fast, cheap, and necessary. They're also the layer attackers find easiest to route around, and the data backs that up.

static-analysis-limits-agent-skills

Field Notes

Three Things That Stall Every Agent Skill Rollout

The scanning is rarely the bottleneck. Across the reviews and rollouts we sit in on, the same three organisational problems show up long before any tool does.

what-stalls-agent-skill-rollouts

Threat Research

When Trusted Isn't: Understanding MCP Impersonation

An MCP server that looks legitimate can still be built to harvest session tokens and tool outputs. Here's what impersonation looks like and what actually stops it.

mcp-impersonation-explained

Threat Research

Beyond the Code: Tool Poisoning and Agent Goal Hijacking Explained

Some of the riskiest agent skill attacks never touch a binary. They target how the agent reads instructions, not how a program executes.

tool-poisoning-agent-goal-hijacking

Threat Research

Backdoors That Outlive the Session: Persistence in Agent Skills

A malicious skill doesn't need to do damage immediately. It just needs to write itself into a shell profile, cron job, or agent config once, and it survives long after the conversation ends.

persistence-backdoors-agent-skills

Threat Research

The curl-Pipe-to-Shell Problem: Remote Script Execution in Agentic AI

Fetching and running a remote script at runtime is a decades-old pattern. In an agent skill, it's one of the fastest paths from install to full compromise.

remote-script-execution-agent-skills

Threat Research

Hidden in Plain Sight: How Obfuscation Hides Data Exfiltration in Agent Skills

Base64 strings and eval-like execution aren't inherently malicious, but they're exactly how attackers keep exfiltration logic invisible to a plain-text code review.

obfuscated-data-exfiltration-agent-skills

Threat Research

Data Thieves in Disguise: How Malicious Agent Skills Harvest Credentials

The most common malicious agent skill archetype doesn't announce itself. It looks like a benign tool while quietly collecting API keys, tokens, and conversation context.

credential-harvesting-agent-skills

Threat Research

Shadow Features: The Malicious Capability Your Skill Scanner Never Sees

A shadow feature is a capability built into a skill but left out of its description. It's why a skill can pass review and still do something no one approved.

shadow-features-malicious-agent-skills

Industry Analysis

What Is an AI Agent Skill, and Why Is It a Supply Chain Risk?

Agent skills give AI assistants the ability to act, not just talk. That power is what makes them the newest link in your software supply chain, and the least inspected one.

what-is-an-ai-agent-skill