Content
65%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The body is an extremely lean redirect stub that delegates the entire skill to an external file outside the bundle. While maximally concise and unambiguous as a directive, it provides no substantive guidance, workflow, or in-bundle references, so the skill is not self-contained and its actual process cannot be evaluated or executed from what is written.
Suggestions
Make the skill self-contained: inline the core workflow steps into SKILL.md, or place the detailed guidance in a file within the bundle (e.g. references/process.md) so the delegated content is actually shipped with the skill.
Add a brief overview of the subagent-driven process (the steps and validation checkpoints) before delegating, so the body is more than a bare redirect and survives the target file being absent.
If the stub pattern is intentional, verify the shared path exists in target environments and note a fallback (e.g. 'if the path is missing, ask the user for the shared skill location') so the redirect cannot silently fail.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is two lines total — a maintenance comment and "Load and follow the shared skill at `~/.agents/skills/subagent-driven-development/SKILL.md`" — with no padding and no explanation of concepts Claude already knows. Every token serves the redirect, matching 'Lean and efficient; every token earns its place'. | 5 / 5 |
Actionability | The single directive is concrete and executable (an exact file path to load), but all substantive guidance is delegated outside the body — 'missing key details' fits, since nothing about how to execute plan tasks is written here. Not 4 because the gap is not minor (100% of the substance is external); not 2 because an exact load-and-follow path is more than a high-level hint. | 3 / 5 |
Workflow Clarity | The redirect action itself is unambiguous, but the skill's described process ("executing implementation plans with independent tasks") is inherently multi-step, and the body contains none of its sequence or validation checkpoints — they are entirely delegated to an external file. The simple-skill exception does not apply because this is not a genuinely simple skill; it hides a complex workflow elsewhere. | 3 / 5 |
Progressive Disclosure | The reference to the shared skill is clearly signaled and only one level deep, but the body is a bare redirect with no overview content, and the referenced path (`~/.agents/skills/...`) lives outside the bundle — no references/, scripts/, or assets/ directories exist. Not 5 because there is no 'clear overview', only indirection to an unverifiable target. | 3 / 5 |
Total | 14 / 20 Passed |