skills/developing-posix-shell-scripts/SKILL.md
Use when writing, reviewing, or refactoring a POSIX shell script (/bin/sh), or when targeting Alpine, BusyBox, or any environment where Bash cannot be assumed.
npx skillsauth add ak1ra-komj/agents-skills developing-posix-shell-scriptsInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
This skill covers any task involving writing, reviewing, or refactoring a POSIX-compliant shell script (/bin/sh).
Bash-specific features are forbidden. See common.md for the full list of bash-isms to avoid.
Audit flags before classifying. Ask for each proposed flag: would a real caller ever pass a different value, or can it be a readonly constant? Beyond that, apply common sense — not everything that could vary should be a flag. Internal paths, fixed timeouts, log levels for non-CLI tools, and similar values are typically hardcoded in practice even if they could theoretically differ. When reviewing an existing script, re-classify from scratch; the current number of flags is not evidence of correct classification.
Simple — all of: < 50 lines of logic, 0–2 genuine flags, no structured logging, no -h/help output, no resource cleanup, not shared across systems/environments.
Complex — any of: ≥ 50 lines, 3+ genuine flags, structured logging, -h/help output, resource cleanup, shared across systems/environments.
When in doubt, prefer Simple.
Always load common.md first, then load the matching document:
development
Use when the user asks to commit changes and/or summarize the current session - or when wrapping up a conversation that involved code or file edits.
development
Use when the user asks to summarize this session or chat — and the conversation is happening in a web chat interface (e.g. ChatGPT Pro) where no file system or terminal access is available.
development
Use when the user asks to commit changes and/or summarize the current session — or when wrapping up a conversation that involved code or file edits.
tools
Use when writing, reviewing, or refactoring AGENTS.md, writing an agent guide, documenting the project for AI agents, or orienting AI coding tools to a new repository.