Content
7%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill reads as an exhaustive reference checklist for Bash scripting rather than an actionable skill document. It is extremely verbose, listing dozens of best practices Claude already knows, while providing almost no executable code examples or concrete workflows. The content would benefit enormously from aggressive condensation, replacing bullet-point advice with a few well-chosen executable examples, and moving the detailed reference material into the referenced sub-skill files.
Suggestions
Reduce the main SKILL.md to under 80 lines by moving detailed sections (Performance, Security, Observability, CI/CD, Modern Features, etc.) into the referenced sub-skill files and keeping only a concise overview with key patterns.
Add 2-3 complete, executable script examples demonstrating the core patterns (strict mode setup, argument parsing with getopts, cleanup traps, structured logging) instead of listing dozens of individual tips.
Replace the vague 4-step Instructions with a concrete workflow: e.g., '1. Start with template (provide template), 2. Run ShellCheck: `shellcheck script.sh`, 3. Fix issues, 4. Run tests: `bats test/`, 5. Only deploy when all pass.'
Remove bullet points that state things Claude already knows (e.g., 'Use `$()` instead of backticks', 'Prefer `printf` over `echo`', basic quoting rules) to dramatically improve token efficiency.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Extremely verbose with extensive bullet-point lists that read like a textbook or reference manual. Much of the content covers things Claude already knows (e.g., 'Use `printf` over `echo`', 'Use `$()` instead of backticks', basic security practices). The skill is over 200 lines of mostly enumerative guidance that could be condensed to a fraction of its size. | 1 / 3 |
Actionability | Despite covering many topics, the skill provides almost no executable code examples. Nearly everything is described in bullet-point prose rather than demonstrated with concrete, copy-paste-ready scripts. The few inline code snippets are fragments, not complete executable examples showing the patterns in context. | 1 / 3 |
Workflow Clarity | The four-step 'Instructions' section is extremely vague ('Define script inputs, outputs, and failure modes', 'Apply strict mode and safe argument parsing') with no concrete workflow, no validation checkpoints, and no feedback loops. For a skill involving potentially destructive operations (file manipulation, CI/CD pipelines), the absence of explicit validation steps is a significant gap. | 1 / 3 |
Progressive Disclosure | The skill references 7 sub-skill files at the bottom, which is a good structural pattern for progressive disclosure. However, no bundle files are provided to verify these exist, and the main SKILL.md itself is a monolithic wall of bullet points that should have been distributed into those sub-skills rather than kept inline. The inline content is far too long for an overview document. | 2 / 3 |
Total | 5 / 12 Passed |