MCP Tool Permissions: How to Limit What an Agent Can Access
Connecting an MCP server can expose many capabilities at once.
Connecting an MCP server can expose many capabilities at once.
Build explicit allowlists, scopes, approvals, and resource boundaries around MCP tools.
Connecting an MCP server can expose a large set of tools.
That is convenient during development but risky in production. The agent should have access to the smallest capability set needed for the current workflow.
If a support agent only needs search_customer and create_ticket, do not expose delete_customer or export_all_customers.
The allowlist should be enforced by the application rather than explained only in a prompt.
Read tools are generally easier to authorize.
Write, delete, send, publish, and financial operations should have stronger controls.
Use separate permissions so a model cannot turn a read-only integration into a mutation interface.
Permission should include the resource boundary.
A user may be allowed to read their own organization but not another customer's records.
Carry tenant, workspace, and user context through every request.
Some operations are too consequential to run automatically.
Require human confirmation for payments, account deletion, publishing, credential changes, or other irreversible actions.
Temporary workflows should not receive permanent permissions.
Where possible, grant access for the duration of the task and revoke it afterward.
Record which tool was requested, who requested it, which policy allowed or denied it, and what resource was targeted.
This creates evidence when something goes wrong.
MCP permissions should be explicit, scoped, temporary where possible, and enforced outside the model.
Discovery makes a tool visible. Authorization decides whether it can actually run.
Source: least-privilege and MCP security principles.
A practical MCP deployment should keep four boundaries explicit: identity, capability, resource, and execution. Identity answers who is acting. Capability answers which tool is allowed. Resource answers which tenant, project, record, or environment can be touched. Execution controls answer how the operation is performed, including timeouts, retries, quotas, and approval requirements.
Do not rely on tool descriptions to enforce these boundaries. Descriptions help the model choose correctly, but the MCP server or trusted backend must reject unauthorized requests. This also protects the system when a prompt injection, buggy client, or compromised workflow sends an unexpected call.
For external dependencies, use bounded timeouts and classify errors. A temporary provider outage may justify a retry, while an authorization failure should stop immediately. For side effects, make duplicate execution safe with idempotency or an operation-status check before retrying.
Test valid and invalid arguments, missing permissions, cross-tenant resource IDs, revoked credentials, unavailable servers, malformed tool results, rate limits, timeouts, and cancellation. Also test what happens when retrieved content contains instructions that attempt to invoke a privileged operation.
For production systems, include integration tests that verify the full chain from authenticated user to MCP client, server policy, downstream service, and sanitized result. A successful local tool call is not enough evidence that the complete security boundary is correct.
MCP becomes valuable when it makes capabilities composable without making authority ambiguous. Keep discovery selective, permissions deterministic, identity explicit, data minimized, and execution observable.
Choose where policy lives before adding more servers. In a small product, the MCP server itself can own authorization and resource checks. In a larger system, a gateway or policy service can provide shared identity, quotas, and audit decisions while each server retains domain-specific validation. What matters is that there is one trusted enforcement path and that clients cannot bypass it.
Keep development capabilities separate from production capabilities. A developer-facing server may expose logs, database inspection, or test data that should never be discoverable by a customer-facing agent. Use separate credentials, registries, and environments where appropriate.
For long-running workflows, persist the workflow state outside the model. If an MCP server disconnects, the system should be able to resume or fail safely without asking the model to reconstruct critical authorization state from conversation history.
Plan for servers disappearing, credentials expiring, downstream APIs returning malformed data, and clients receiving stale tool definitions. A discovery failure should not silently fall back to a broader capability. A permission failure should not be converted into a generic retry. A timeout should not automatically mean that a side effect did not happen.
Use health checks, bounded retries, circuit breakers for unstable dependencies, and clear degraded states. When a tool becomes unavailable, the agent should know that capability is unavailable rather than inventing a result.
Before production, ask whether every exposed tool has a clear owner, a documented purpose, a resource boundary, a permission model, a timeout, an error contract, and an audit trail. Remove tools that are unused or whose authority cannot be explained precisely.
The goal is not the largest MCP tool catalog. The goal is a small set of capabilities that an agent can use predictably and that engineers can explain during a security review.
Start with one read-only server and a small allowlist of tools. Measure discovery, execution latency, denied calls, errors, and successful task completion. Once the read path is stable, add one carefully scoped write operation with an approval requirement. Only then consider broader integrations.
This staged approach makes failures easier to isolate and keeps the blast radius small. It also creates useful evidence for deciding which tools deserve permanent access. Remove capabilities that provide little value relative to their security and maintenance cost.
Can a tool access a resource outside the current tenant? Can retrieved content influence a privileged operation? Can a retry duplicate a side effect? Can an expired credential remain usable? Can a client discover a production-only capability? Can an error reveal sensitive data? Can one compromised server reach unrelated internal systems?
If any answer is unclear, the integration is not finished. Make the boundary explicit in code, tests, and documentation before expanding the tool set.
Community
0 comments
React to this article
Trending now
Written by
Kirtesh Admute
Founder
Kirtesh Admute is the founder of IndieFounder, a platform for founders, builders, and people curious about technology. He writes about AI, startups, software, product building, and the lessons that come from building in public.
See an issue with this story?
