# MCP tool descriptions and schemas risks

URL: https://mcpservercybersecurity.com/mcp-risk-register/tool-descriptions-and-schemas/
Updated: 2026-11-16

Six risks sit in what the model reads: poisoned descriptions, shadowing across servers, hostile schemas and return values, trusted annotations, prompt injection in tool results and server-supplied schemas that burden clients. Each row cites OWASP, the MCP spec or published research, with severity from one written rule, read on 16 Nov 2026.

## What this dimension covers

What the model reads: descriptions, schemas, annotations and tool results. The register holds 6 rows here (R-23, R-24, R-25, R-26, R-27, R-28). Each row states what the risk is, its impact, concrete controls and at least one primary source.

Sources: [OWASP MCP Security Cheat Sheet](https://cheatsheetseries.owasp.org/cheatsheets/MCP_Security_Cheat_Sheet.html); [When an MCP Tool Asks for Your GitHub Token, Knostic](https://www.knostic.ai/blog/when-an-mcp-tool-asks-for-your-github-token); [AI tool poisoning through MCP tool descriptions, CrowdStrike](https://www.crowdstrike.com/en-us/blog/ai-tool-poisoning/).

## Highest-severity rows

Rows that meet the high rule under the written rule are: R-23 Poisoned tool description; R-27 Prompt injection through tool results.

- R-23: The agent takes an action the user did not request, such as reading a credential.
- R-27: An approved server can still cause unapproved actions through the content it returns.

Sources: [OWASP MCP Security Cheat Sheet](https://cheatsheetseries.owasp.org/cheatsheets/MCP_Security_Cheat_Sheet.html); [When an MCP Tool Asks for Your GitHub Token, Knostic](https://www.knostic.ai/blog/when-an-mcp-tool-asks-for-your-github-token); [AI tool poisoning through MCP tool descriptions, CrowdStrike](https://www.crowdstrike.com/en-us/blog/ai-tool-poisoning/); [MCP Tools, 2026-07-28 revision](https://modelcontextprotocol.io/specification/2026-07-28/server/tools).

## First controls to apply

Start with the controls on the high rows, then work down. Controls name settings only where a primary source documents them; check each against the version your clients run.

- R-23: Review the full text of every tool description and parameter field before a server is approved.
- R-24: Treat each server as its own security domain and keep sensitive servers apart from general-purpose ones.
- R-25: Inspect every field of the definition, including inputSchema and outputSchema.
- R-26: Treat annotations as hints only and decide approval from your own policy per tool.
- R-27: Treat every tool response as untrusted data, including responses from approved servers.
- R-28: Evaluate server-supplied patterns with a linear-time matcher and limits.

Sources: [OWASP MCP Security Cheat Sheet](https://cheatsheetseries.owasp.org/cheatsheets/MCP_Security_Cheat_Sheet.html); [When an MCP Tool Asks for Your GitHub Token, Knostic](https://www.knostic.ai/blog/when-an-mcp-tool-asks-for-your-github-token); [AI tool poisoning through MCP tool descriptions, CrowdStrike](https://www.crowdstrike.com/en-us/blog/ai-tool-poisoning/); [MCP Tools, 2026-07-28 revision](https://modelcontextprotocol.io/specification/2026-07-28/server/tools).

## Tool descriptions and schemas: 6 items

### R-23: Poisoned tool description

Instructions hidden in a tool description steer the model, which reads the description to decide how to call the tool. Knostic described a package on 12 Nov 2026 whose description tells the agent to fetch a GitHub token without asking the user.

**Impact:** The agent takes an action the user did not request, such as reading a credential.

**Controls:**

- Review the full text of every tool description and parameter field before a server is approved.
- Hash and pin the reviewed definitions and alert on change.
- Scan descriptions with a tool built for the purpose (the OWASP cheat sheet names mcp-scan) and keep a human approval step for sensitive calls.

Severity: high

Sources: [OWASP MCP Security Cheat Sheet](https://cheatsheetseries.owasp.org/cheatsheets/MCP_Security_Cheat_Sheet.html) (read 2026-11-16); [When an MCP Tool Asks for Your GitHub Token, Knostic](https://www.knostic.ai/blog/when-an-mcp-tool-asks-for-your-github-token) (read 2026-11-16); [AI tool poisoning through MCP tool descriptions, CrowdStrike](https://www.crowdstrike.com/en-us/blog/ai-tool-poisoning/) (read 2026-11-16).

### R-24: Tool shadowing across servers

One server's tool text manipulates how the agent uses tools from another, trusted server, because the model sees all connected descriptions at once.

**Impact:** A low-trust server can alter how a high-trust one is used, including which data is sent where.

**Controls:**

- Treat each server as its own security domain and keep sensitive servers apart from general-purpose ones.
- Monitor for cross-server data flows, such as one server's credentials appearing in another's calls.
- Limit how many servers a single agent session may load.

Severity: medium

Sources: [OWASP MCP Security Cheat Sheet](https://cheatsheetseries.owasp.org/cheatsheets/MCP_Security_Cheat_Sheet.html) (read 2026-11-16).

### R-25: Hostile content in schemas and return values

Instructions can sit in parameter names, types, defaults and the values a tool returns, not only in the description field.

**Impact:** Review that checks only the description misses the other fields the model reads.

**Controls:**

- Inspect every field of the definition, including inputSchema and outputSchema.
- Set additionalProperties to false and use pattern constraints on string fields.
- Validate and sanitize tool outputs before they re-enter model context.

Severity: medium

Sources: [OWASP MCP Security Cheat Sheet](https://cheatsheetseries.owasp.org/cheatsheets/MCP_Security_Cheat_Sheet.html) (read 2026-11-16); [MCP Tools, 2026-07-28 revision](https://modelcontextprotocol.io/specification/2026-07-28/server/tools) (read 2026-11-16).

### R-26: Annotations trusted as enforcement

A client lets a server's readOnlyHint or destructiveHint decide how much approval to ask for. The spec says clients must consider annotations untrusted unless they come from trusted servers.

**Impact:** A flipped hint can change approval behavior with no visible change to the description.

**Controls:**

- Treat annotations as hints only and decide approval from your own policy per tool.
- Include annotations in the pinned hash of each tool definition.
- Require confirmation for sensitive reads, writes and irreversible actions whatever the label says.

Severity: medium

Sources: [MCP Tools, 2026-07-28 revision](https://modelcontextprotocol.io/specification/2026-07-28/server/tools) (read 2026-11-16); [OWASP MCP Security Cheat Sheet](https://cheatsheetseries.owasp.org/cheatsheets/MCP_Security_Cheat_Sheet.html) (read 2026-11-16).

### R-27: Prompt injection through tool results

A tool returns text, such as a web page or ticket, that contains instructions the model then follows.

**Impact:** An approved server can still cause unapproved actions through the content it returns.

**Controls:**

- Treat every tool response as untrusted data, including responses from approved servers.
- Enforce authorization and validate follow-up tool calls in application code, regardless of the model's interpretation.
- Extract structured fields from retrieval tools instead of passing raw pages.

Severity: high

Sources: [OWASP MCP Security Cheat Sheet](https://cheatsheetseries.owasp.org/cheatsheets/MCP_Security_Cheat_Sheet.html) (read 2026-11-16).

### R-28: Server-supplied schemas that exhaust or confuse clients

A client evaluates patterns or fetches references in a schema a server supplies. Inspector 2.14.0 now checks schema patterns in bounded time, and Python SDK advisory GHSA-rwrf-2pqf-9j8j describes a client fetching server-chosen $ref URLs.

**Impact:** A hostile server can slow or break a client, or cause it to make requests it should not.

**Controls:**

- Evaluate server-supplied patterns with a linear-time matcher and limits.
- Do not resolve remote references from schemas supplied by untrusted servers.
- Keep MCP clients and the Inspector on current releases.

Severity: low

Sources: [MCP Inspector 2.14.0 release notes](https://github.com/modelcontextprotocol/inspector/releases/tag/2.14.0) (read 2026-11-16); [GHSA-rwrf-2pqf-9j8j: MCP Python SDK server-chosen $ref URLs](https://github.com/advisories/GHSA-rwrf-2pqf-9j8j) (read 2026-11-16).

## Related pages

- [The full MCP Risk Register](/mcp-risk-register/)
- [Transport and network](/mcp-risk-register/transport-and-network/)
- [Execution and filesystem](/mcp-risk-register/execution-and-filesystem/)
- [Method](/mcp-risk-register/methodology/)
- [MCP Security Checklist](/mcp-security-checklist/)
