rm -rf "$DIR/"* deletes everything when DIR is empty, and one eval "$cmd" lets anyone who controls the text run commands. This skill finds those and eight other shell script mistakes before you run or ship your scripts, and tells you how to fix each one.

Your agent runs a scanner on .sh, .bash and .zsh files, scripts with a shell shebang (such as bin/deploy and git hooks) and shell startup files. It skips comments, knows single and double quotes, treats heredoc bodies as data, and knows the usual guards (${DIR:?}, [ -n "$DIR" ], || exit, set -e). It never runs the script. It returns BLOCK, REVIEW, or NOT FLAGGED, with the file, line, reason, and a safer approach. Secrets are never printed.

What it checks

BLOCK (fix before shipping)

• S01: eval of a variable or argument

• S02: passwords and tokens in assignments, curl -u, Authorization headers, mysql -p or sshpass

• S03: rm -r on "$DIR/"* or $DIR/ when DIR may be empty, or rm -rf /

REVIEW (check before release)

• S11: curl or wget piped into a shell

• S12: unquoted variables in rm, mv, cp, chmod and chown

• S13: curl -k, wget --no-check-certificate or StrictHostKeyChecking=no

• S14: chmod 777, o+w or umask 000

• S15: writes to fixed file names in /tmp

• S16: set -x in a script that uses secrets

• S17: cd without a check before rm -r

What's included

• The scanner (Python 3.9+, standard library only)

• A rules guide with the reason and safer approach for each rule

• A risky and a safe example deploy script, and the real report the scanner produces on the risky one

Limits

It reads each script line by line: sourced files, functions from other files and commands built across several variables are not followed, and shell code inside Dockerfiles, CI YAML or Makefiles is not read. NOT FLAGGED is not a guarantee.

Contents

One ZIP file with 8 files: SKILL.md, the scanner in scripts/ (5 Python files), references/rules.md (every rule with its reason and safer approach), and examples/ (2 example files plus report.md, the real scanner output on the risky example).

How to use: This is an agent skill: a folder with a SKILL.md file and Python scripts that an AI coding agent such as Claude Code, Cursor, or Codex CLI can run. Unzip it into your agent's skills folder (for Claude Code: ~/.claude/skills/), then ask your agent to check your project. The scripts also run on their own with Python 3.9 or newer, with nothing to install. They read your files only: no network access, and nothing is sent anywhere.