Shield KYA vs Snyk Agent Scan
Short answer: Snyk Agent Scan checks the agent components on a machine for security risks. Shield KYA records the tool calls that your agents run. They do different jobs, and you can use both.
Shield KYA is an open-source audit trail and policy gate for AI coding agents, for teams who need to see and stop what agents run on developer machines.
Comparison
| Shield KYA | Snyk Agent Scan | |
|---|---|---|
| What it is | An open-source audit trail and policy gate for AI coding agents | "Security scanner for AI agents, MCP servers and agent skills" |
| What it looks at | The tool calls that a coding agent sends through its hook or MCP gate | MCP configurations, agent tools and skills on the machine |
| What you get | A local trail, a local dashboard and a receipt of what ran. The hosted workspace (preview) adds human approval. | Findings across 14 security risks, such as prompt injection, sensitive data handling and malware payloads hidden in natural language |
| Blocking | The open-source CLI is advisory and blocks only a short never-list. The hosted workspace (preview) enforces policy and fails closed. | It is a scanner. Its page describes a scan mode and a background mode. |
| Agents named | Claude Code, Codex, Gemini CLI, GitHub Copilot CLI, Grok, Kimi Code (hook) and Cursor (MCP) | Claude, Cursor, GitHub Copilot, Windsurf, Gemini CLI, Amp, Amazon Q and others |
| Account | None for the open-source CLI | A Snyk API token (SNYK_TOKEN) |
| Data that leaves the machine | The trail stays on your machine unless you connect a hosted workspace or share a report. Anonymous usage stats are opt-in and off by default. | Agent application details, tool names and descriptions, and skill content go to Snyk for analysis. Snyk redacts secrets before it sends them. |
| Install | npm i -g @shield-agent/kya@latest && kya start |
uv or a standalone binary from GitHub Releases |
| License | MIT | Apache-2.0 |
Trade-offs
- Use Snyk Agent Scan to find a risky MCP server or skill before an agent uses it.
- Use Shield KYA to see what each agent ran, and to stop the calls on its never-list.
- Snyk says that scanning MCP configurations can execute commands or make network requests. Read its consent prompts.
- The Shield KYA open-source CLI is advisory. It is not your only barrier. Keep the native deny rules of your agents.
What Shield KYA does
- The open-source CLI (
@shield-agent/kya, MIT license) records the tool calls of your coding agents in a local trail and a local dashboard. - The CLI is advisory and fails open on internal error. Its local hook blocks only a short never-list and a guard pack: destructive shell commands, secret reads and exfiltration. It records all other calls and does not block them.
- The hosted workspace (preview) enforces policy, holds actions for human approval in its console, and fails closed.
- A pre-tool hook sees each tool call that the agent sends through it in Claude Code, Codex, Gemini CLI, GitHub Copilot CLI, Grok and Kimi Code. In Cursor, Shield KYA sees the calls that go through its MCP gate. It does not see agents that never call it.
- It does not screen prompts or model output. Use a guardrails product for that.
FAQ
Can I run Shield KYA and Snyk Agent Scan together?
Yes. The scan checks configuration. Shield KYA records and checks tool calls at run time. They do not replace each other.
Does Shield KYA scan MCP servers for poisoned tool descriptions?
No. kya discover (preview) lists the AI agents, IDEs, runtimes and MCP servers on a machine. It does not check tool descriptions for prompt injection. Snyk Agent Scan does that.
Does Shield KYA need an account?
No. The open-source CLI runs on your machine without an account. The hosted workspace needs an account.