Content
50%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 provides a clear procedural sequence with genuinely useful guardrails around the Git remote check, but its core value — creating dependency-ordered issues — is the least specified step, and the duplicated hook-checking blocks consume significant tokens. Adding issue-format guidance and a post-creation verification step would address the weakest areas.
Suggestions
Specify the issue format concretely: show an example issue title/body derived from a task and state how dependencies are encoded (e.g. blocked-by references or labels).
Factor the duplicated hook-checking procedure into a single shared section or a reference file, noting only the key difference (hooks.before_taskstoissues vs hooks.after_taskstoissues).
Add a verification checkpoint after issue creation (e.g. confirm created issue URLs/count against the task list) to satisfy the batch-operation validation requirement.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly efficient procedural instruction with no concept over-explanation, but the ~25-line hook-checking procedure is duplicated nearly verbatim across the Pre-Execution and Post-Execution Checks sections (enabled filtering, condition handling, dot-to-hyphen conversion, output templates). Anchor 3 rather than 2 because the duplicated content is actionable instruction, not padding or explanation of known concepts, though it could clearly be tightened. | 3 / 5 |
Actionability | Concrete elements exist ("Run .specify/scripts/powershell/check-prerequisites.ps1 -Json -RequireTasks -IncludeTasks", "git config --get remote.origin.url", exact hook output templates), but the core step — "use the GitHub MCP server to create a new issue" — is under-specified: no issue title/body structure, no example of how dependencies are represented, no label usage. Anchor 3 rather than 4 because a key executable detail for the main action is missing. | 3 / 5 |
Workflow Clarity | The sequence (hook checks → prerequisites → extract tasks → verify remote → create issues → post hooks) is present with an explicit remote-URL guard ("UNDER NO CIRCUMSTANCES EVER CREATE ISSUES IN REPOSITORIES THAT DO NOT MATCH THE REMOTE URL"), but this is a batch operation with no verification that issues were created correctly and no error-recovery loop mid-batch. The rubric's batch-operation cap applies, holding workflow clarity at 3 rather than 4. | 3 / 5 |
Progressive Disclosure | Sections are well-organized (User Input, Pre-Execution Checks, Outline, Post-Execution Checks) with no nested references, but the ~55 lines of hook machinery are inlined and duplicated in the main body rather than split into a reference file, and the skill has no bundle files at all. Anchor 3 rather than 4 because content that clearly belongs in a separate file is inline and repeated. | 3 / 5 |
Total | 12 / 20 Passed |