Menu

1 / 2

Your AI database agent should not approve its own writes

DEV Community·Mads Hansen·6 months ago
#m4vdF9wW
#mcp#database#ai#security#approval
Reading 0:00
15s threshold

The riskiest AI database workflow is not a bad SELECT .

It is a write that looks reasonable.

Update the customer status.

Fix the subscription record.

Mark these invoices as reviewed.

Some writes are legitimate.

That does not mean the model should approve them.

The model should propose, not decide

An AI agent can help prepare a change:

inspect context

draft the SQL or API call

explain the expected side effect

identify related records

produce a dry-run summary

But approval should live outside the model loop.

If the same system that generated the change also decides it is safe, the approval gate is mostly theater.

Show the diff before approval

A useful approval request should include:

exact operation

affected entity IDs

before and after values

estimated row count

policy rule that requires approval

rollback or compensation path

audit identifier

Approval is not a substitute for scope.

Writes still need tenant scope, role separation, type validation, and deterministic execution.

Longer…

Continue reading — create a free account

Join HashtagPLUS to read full articles, follow hashtags, vote, and join the conversation.