Production deploy script safety
You want an AI agent to write a Bash deploy script that pulls the latest build, backs up the current release, swaps in the new one, and restarts a systemd service. Write the prompt so the script is safe to run in production. What constraints make a Bash script trustworthy, and what does a quick 'write a deploy script' prompt get wrong?
Implement
audit_deploy_script(lines: list[str]) → list[str]Examples
in
[["#!/usr/bin/env bash","set -euo pipefail","cp -r \"$RELEASE_DIR\" \"$BACKUP_DIR\"","systemctl restart \"$SERVICE\" || rollback"]]out[]in
[["#!/bin/bash","cd $DEPLOY_DIR","rm -rf $RELEASE_DIR","systemctl restart myservice"]]out["missing_pipefail","missing_rollback","unquoted:2","unquoted:3"]in
[["set -euo pipefail","trap rollback ERR","tar -xzf \"$ARTIFACT\" -C $TARGET"]]out["unquoted:3"]What a strong answer looks like
Treat the AI’s output as a draft to verify, not an answer to trust. Name the specific flaw and the input that triggers it, say how you’d catch it (tests, edge cases, reading critically), and how you’d re-prompt or decompose to get it right.
0:00 of about 12 min
Vibe & agentic: describe the solution in plain language (or narrate it) and the coach grades your approach.
Which questions mattered is sealed until you submit. Telling you now would just be handing over the edge cases.
Run or narrate your approach, then ask the coach.