loadout shell.instructions.md

A set of rules for writing production Bash shell scripts, including command-line tools, automation, Git hooks, and script wrappers. Bash is a program used to run commands and automate tasks on many computers.

In plain words
What is it for?
Writing or reviewing Bash scripts for automation, CI/CD pipelines, Git hooks, command-line utilities, and tools that call other programs.
Why use it?
Shell scripts can fail silently, mishandle input, or behave differently across environments. The instructions provide consistent practices for structure, variables, errors, and documentation.

Instructions file for GitHub Copilot

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add instructions/stefaniuk/loadout/shell
Clone the repo
git clone --depth 1 https://github.com/stefaniuk/loadout

Made for: GitHub Copilot.

Per session 6,830 This file is loaded in full into every session.
When invoked 6,830 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.06830 $0.06830
Opus 5 $0.03415 $0.03415
Sonnet 5 $0.01366 $0.01366
Haiku 4.5 $0.00683 $0.00683

Measured yesterday against content hash 4909c1a6456e, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

loadout shell.instructions.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.github/instructions/shell.instructions.md · 785 lines

How it starts

The opening of the file, as written. The whole thing — 785 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Shell Script Engineering Instructions (Bash, CLI wrappers, automation) 🐚

These instructions define the default engineering approach for writing production-grade shell scripts in Bash.

They must remain applicable to:

  • Standalone CLI utility scripts
  • Git hook scripts
  • CI/CD pipeline scripts
  • Tool wrapper scripts (native/Docker dual execution)
  • Library scripts sourced by other scripts
  • Report generation and automation scripts

They are non-negotiable unless an exception is explicitly documented (with rationale and expiry) in an ADR/decision record.

Cross-references. For Makefile orchestration conventions that call shell scripts, see makefile.instructions.md. This file focuses exclusively on shell script patterns.

Identifier scheme. Every normative rule carries a unique tag in the form [SH-<prefix>-NNN], where the prefix maps to the containing section (for example QR for Quick Reference, HDR for Header, FN for Functions, VAR for Variables, ERR for Error Handling, DOC for Documentation). Use these identifiers when referencing, planning, or validating requirements.


0. Quick reference (apply first) 🧠

This section exists so humans and AI assistants can reliably apply the most important rules even when context is tight.

  • [SH-QR-001] Shebang and safety: start with #!/bin/bash and set -euo pipefail ([SH-HDR-001], [SH-HDR-003]).
  • [SH-QR-002] Header structure: include WARNING comment, safety options, description block, and section dividers ([SH-HDR-002], [SH-HDR-004]).
  • [SH-QR-003] main() entry point: use a main() function as the script entry point, called at the bottom of the file ([SH-FN-001]).
  • [SH-QR-004] Kebab-case functions: name functions using kebab-case (lowercase with hyphens) ([SH-FN-002]).
  • [SH-QR-005] Document all functions: provide comment blocks describing purpose and arguments ([SH-DOC-001]–[SH-DOC-004]).
  • [SH-QR-006] Native/Docker dual execution: scripts should run tools natively if installed, or fall back to Docker ([SH-EXEC-001]–[SH-EXEC-005]).
  • [SH-QR-007] is-arg-true() helper: use a standard boolean-parsing helper for environment flags ([SH-VAR-006]).
  • [SH-QR-008] VERBOSE toggle: support VERBOSE=true to enable set -x for debugging ([SH-DBG-001]–[SH-DBG-003]).
  • [SH-QR-009] Exit 0 at end: scripts must end with exit 0 after successful execution ([SH-ERR-005]).
  • [SH-QR-010] Test suite patterns: use Arrange-Act-Assert structure, setup/teardown functions, and summary reporting ([SH-TST-001]–[SH-TST-015]).
  • [SH-QR-011] Avoid common anti-patterns: unquoted variables, cd without error handling, hardcoded paths (§12).
  • [SH-QR-012] Explicit returns: finish every function with an explicit return <status> so intent and requirements stay aligned ([SH-FN-008]).

Read the full file on GitHub · 785 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 785 lines · 6,830 tokens per session scan A 4909c1a6456e

Subscribe to this mod's changes

loadout shell.instructions.md is an instructions file published in the GitHub repository stefaniuk/loadout (1 stars, last pushed 3d ago), licensed MIT. It adds 6,830 tokens to every session, about $0.0341 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.