// GLOSSARY -- MCP & TOOL INFRASTRUCTURE

What is a Tool Risk Category?

1 min read Updated

A classification label (Read, Write, Execute, Destructive, Financial) assigned to an MCP tool based on its potential impact, used to enforce graduated approval and policy controls.

WHY IT MATTERS

Not all tools carry equal risk. Reading a calendar event is fundamentally different from deleting a database table or initiating a payment. Without classification, every tool gets the same trust level — either everything is allowed or everything requires approval.

Risk categories enable graduated enforcement. Read tools can be auto-approved. Write tools get rate limits. Destructive tools require human confirmation. Financial tools need budget checks. Classification is the foundation of proportionate access control.

See tool risk category working in your own stack — route your MCP servers through PolicyLayer and every tool call is checked against policy before it runs.

GOVERN YOUR MCP SERVERS →

Enforced before the call runs. Nothing to install.

HOW POLICYLAYER USES THIS

PolicyLayer classifies 18,000+ MCP tools across 3,100+ servers into five risk categories using deterministic pattern matching on tool names, descriptions, and input schemas. This classification powers PolicyLayer's policy engine.

See the full risk breakdown →

IN THE CATALOGUE

Across the 530,019 classified tools in PolicyLayer's published scan reports, this is how the risk surface actually distributes.

36,613 destructive or financial tools (6.9% of all tools)
46,956 execute tools — run code or trigger actions
10,538 servers expose at least one critical tool

See the full risk breakdown →

FREQUENTLY ASKED QUESTIONS

What are the five categories?
Read (data retrieval), Write (create or modify data), Execute (run code or commands), Destructive (delete or irreversible actions), Financial (payments, transfers, billing).
How is classification done?
Deterministic pattern matching on tool names, descriptions, and input schemas — no LLM in the loop, so results are reproducible and auditable.
Can I override a classification?
Yes. PolicyLayer policies let you override the default category for any tool. If your use case treats a 'write' tool as low-risk, you can configure accordingly.

FURTHER READING

Take your agents live. Without losing control.

Route your MCP traffic through PolicyLayer. Every tool call is checked against your policy before it runs: allow, deny, or require approval. Per-identity grants. Full audit log. Live in minutes.

GOVERN YOUR MCP SERVERS →

Instant setup, no code required.

46,500+ MCP servers and 515,000+ tools scanned and risk-classified.

// GET IN TOUCH

Have a question or want to learn more? Send us a message.

Message sent.

We'll get back to you soon.