Urgent.News

What's breaking now, across thousands of outlets.

Tech

Which Sheet Automations Get a Human Approval Step, and Which Just Run?

Every Apps Script job we build for a small team ends with the same argument: should this run straight through, or should it stop and wait for a person? Our current line is roughly this: Runs straight through: reads, summaries, formatting, and anything that only writes inside the same spreadsheet and can be redone by running it again. Writes a plan first, then waits: anything that deletes, sends…

Every Apps Script job built for a small team faces a critical decision: should it proceed uninterrupted, or should it halt and await human intervention? The consensus within the team is as follows: jobs that read, summarize, format, and solely modify data within the same spreadsheet without resulting in irreversible changes are allowed to run automatically.

Conversely, jobs that delete data, send information outside the company, or handle financial transactions must first write a plan and then wait for human review before executing the final steps.

To illustrate this point, the team conducted a dry run for Workspace offboarding. In this scenario, the reviewer is required to input the leaver's email address before the apply step can be executed. This is due to the ease with which a checkbox can be ticked without proper attention to the plan. The downside of this approval step is clear: every time the person who approves is away, the job idles at this point. Furthermore, a plan that goes unread essentially becomes an inefficient method of repeating the same mistakes.

For those who manage automations for small teams, the question arises: how do you determine the line between "just run" and "ask first"? Additionally, who steps in to approve when the usual approver is unavailable? There have been instances where teams have removed approval steps altogether because they deemed the approvers as unengaged and non-reading.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dev.to →

More in Tech

Antigravity and Codex security: stopping secret leaks with hooks and sandboxes

Originally published at ReclaimLLM . How Google Antigravity and OpenAI Codex agents read, print, and send secrets; what hooks and sandboxes cover; and where vendor docs leave failure modes…

  • Antigravity and Codex provide hooks for denying tool calls.
  • Sandboxes and approval boundaries restrict shell command capabilities.
  • Combine pre-tool hooks, network limits, and session capture for secret protection.

Building a Work-Hours Calculator? Time Arithmetic Is the Trap

If you have ever built a timesheet tool, a shift planner, or a payroll helper, you know the feature looks trivial until you write the first test.

  • Store times as minutes since midnight to avoid floating-point rounding errors
  • Perform calculations using integers, subtracting breaks before rounding, only round at end
  • Use modular arithmetic for overnight shifts, wrap time difference around 24-hour day

More from Sunday 1 November →