Content
72%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The skill body is a well-structured, actionable dispatcher with excellent progressive disclosure and concrete routing guidance. Its weaknesses are the bulky undifferentiated file catalog hurting conciseness and the absence of an explicit checkpoint validating that the right prompt was chosen.
Suggestions
Trim the raw ~58-file catalog to only the routable files, or relocate the full list to a reference file and keep the overview focused on the routing table.
Add an explicit validation step for the routing decision, e.g., 'Before loading, confirm the task type matches exactly one prompt; if ambiguous, ask a clarifying question and do not proceed.'
Annotate the catalog entries that are not in the routing table (e.g., note they are reference-only) so the reader knows which files are loadable on demand versus core routes.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly lean and avoids concept padding, but the ~58-path raw bulleted catalog includes many unrouted files that don't all earn their place in the overview and could be tightened or moved to a reference. | 2 / 3 |
Actionability | It provides concrete, loadable file paths and an explicit task-to-file routing table ('Design workflows from scratch via interview: `.github/aw/designer.md`'), plus a concrete overlay-loading directive with `@.github/aw/instructions.md`. | 3 / 3 |
Workflow Clarity | The dispatch sequence (identify task type -> load matching prompt -> follow it) is clear, but there is no explicit validation checkpoint confirming the correct prompt was selected beyond a soft 'ask a clarifying question' fallback. | 2 / 3 |
Progressive Disclosure | It is a well-organized overview pointing to one-level-deep references with a clearly signaled routing section mapping each task type to its file, enabling easy navigation. | 3 / 3 |
Total | 10 / 12 Passed |