Developers complain that agents wait for approval on every small edit. What is the right change?
Why A is the answer
Classify actions by risk and require approval only where it reduces risk
Dropping approvals that do not reduce risk keeps velocity, while irreversible or sensitive changes keep explicit authorization.
In detail
- Approvals should go where they reduce risk.
- Classify actions, then let low-risk work flow.
- Review the classification as you learn.
Use case
An illustrative scenario, not a real incident.
Developers wait for approval on each typo fix. Classifying by risk means small edits flow, while a change to authentication needs a person.
See it in different ways
flowchart TD
A{"How risky is the action?"} -->|"Low and reversible"| B["Agent acts, people review later"]
A -->|"Medium"| C["Agent proposes in a pull request"]
A -->|"High"| D["Human approval before acting"]sequenceDiagram
participant A as Agent
participant V as Validation
participant R as Repository
A->>V: Structured plan
alt Plan passes
V->>A: Go ahead
A->>R: Makes changes in a branch
else Plan fails
V-->>A: Fix the plan first
endflowchart TD
Q{"How risky is the action?"}
Q -->|"Low"| A0["Runs with logging"]
Q -->|"High"| A1["Needs approval"]Key point: Approve where it reduces risk.
Why the other options are wrong
- B. Remove all approvals so agents can move at full speed
- Removing all approvals leaves risky actions unchecked.
- C. Add a second approver so reviews are spread across the team
- A second approver adds more waiting.
- D. Stop using agents for small edits and keep them for large ones
- Small edits are often a good fit for agents.