Lakera alternatives for coding-agent security
Short answer: Lakera Guard, now documented as Check Point AI Guardrails, screens the content that goes into an LLM and the LLM output. If your risk is what coding agents run on developer machines, also look at tools that act on tool calls: the native agent controls, Microsoft Agent Governance Toolkit, Snyk Agent Scan and Shield KYA.
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
| Option | What it acts on | How it works |
|---|---|---|
| Check Point AI Guardrails (Lakera Guard) | User and external content that goes into an LLM, and the LLM output | Your application sends a POST request to the guard endpoint for each LLM interaction. It detects prompt attacks, data leakage, content violations and unknown links. |
| Native controls in Claude Code, Cursor and Codex | The tool calls, commands and file access of that one agent | Permission rules, approval prompts, hooks and sandboxes inside each agent. Admins can manage some settings centrally. |
| Microsoft Agent Governance Toolkit | Tool calls, message sends and delegations | Intercepts each action in deterministic code, with a fail-closed policy runtime and an audit trail. Open source (MIT), public preview. |
| Snyk Agent Scan | MCP configurations, agent tools and skills on a machine | A scanner that detects 14 security risks. Needs a Snyk API token. |
| Shield KYA | Each tool call of your coding agents | Records each call in one local trail. The open-source local hook blocks only a short never-list. The hosted workspace (preview) enforces policy, holds actions for human approval and fails closed. |
Trade-offs
- Keep Lakera Guard if you build an LLM application and need to screen prompts and output. Shield KYA does not screen prompts or model output.
- The Lakera site describes runtime protection for AI applications and names agents and MCPs. Check its current pages for coding-agent support.
- Native controls block well inside one agent. They do not give one record across agents.
- Microsoft Agent Governance Toolkit blocks tool calls and fails closed. Shield KYA's one command shows what your coding agents did, with opt-in human approval and signed certify, locally or hosted.
- The Shield KYA open-source CLI is advisory. It is not your only barrier.
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
Is Shield KYA a replacement for Lakera Guard?
No. They act on different things. Lakera Guard screens LLM content. Shield KYA records and checks the tool calls of coding agents. A team can use both.
Is Lakera Guard now Check Point AI Guardrails?
The Lakera docs say: "Check Point AI Guardrails screens user and external content going into LLMs and the resulting output".
Which option blocks an agent action before it runs?
The native controls of each agent and Microsoft Agent Governance Toolkit. The Shield KYA hosted workspace (preview) also enforces policy. The Shield KYA open-source CLI blocks only its never-list.