Content
100%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A well-engineered instruction skill: concise convention statements backed by executable bats test patterns, a useful anti-patterns table, honest limitations, and a correctly split one-level reference file. It respects Claude's competence while providing concrete, copy-paste-ready assertions.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean: short convention sections each followed by a tight executable bats snippet, an anti-patterns table, and a brief limitations list. The clig blockquotes are brief authoritative citations rather than pedagogical padding, so every token earns its place; not a 2 because there is no teach-the-basics bloat. | 3 / 3 |
Actionability | Each convention ships copy-paste-ready bats assertions (e.g. "run --separate-stderr mycli list", "[ \"$status\" -eq 2 ]", "echo \"$output\" | jq -e ...") plus concrete good/bad shell transcripts, matching the 'fully executable code/commands; copy-paste ready' anchor. | 3 / 3 |
Workflow Clarity | Sequencing is explicit via the "When to use" section and per-convention test patterns; for a convention/contract skill (single-purpose, non-destructive) the guidance is unambiguous. Per the simple-skills scoring note, an unambiguous single-purpose skill can score 3 without destructive-op validation checkpoints. | 3 / 3 |
Progressive Disclosure | SKILL.md is an overview with conventions 1-4 inline and conventions 5-8 clearly signaled via a one-level-deep reference to references/conventions-5-8.md (a real file). Organization is clean with easy navigation, matching the 'clear overview with well-signaled one-level-deep references' anchor. | 3 / 3 |
Total | 12 / 12 Passed |