Content
85%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A well-structured, highly actionable skill body with clear sequenced workflows, explicit validation for the batch sync operation, and accurate progressive disclosure that matches the real bundle. The main weakness is mild redundancy in the introductory prose, which costs it the top conciseness score.
Suggestions
Remove the duplicated phrasing: "turning query data, source notes, and page evidence into structured decisions instead of improvised prompts" appears almost verbatim in both the opening paragraph and the FLOW overview section — keep it in one place only.
Tighten the conceptual intro (lines 27-35): the explanation of what FLOW is and why local-SEO prompts are excluded can be condensed to a single sentence each, since the commands and reference sections already convey the surface area.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Largely efficient and actionable, but the phrase "turning query data, source notes, and page evidence into structured decisions instead of improvised prompts" appears nearly verbatim in both the opening (lines 21-23) and the FLOW overview (lines 28-30), and the introductory prose explaining what FLOW is could be tightened; not every token earns its place. | 2 / 3 |
Actionability | Concrete, executable guidance throughout: exact sub-commands, specific file paths to read per stage, a defined 2-to-3 prompt selection rule with stated criteria, exact sync invocations with flags (--dry-run, --ref <sha>), and a copy-paste-ready attribution block. | 3 / 3 |
Workflow Clarity | Each sub-command has a clearly numbered sequence, and the batch sync operation includes explicit validation (lockfile drift reported every run, atomic writes, path-traversal guard) plus an error-handling table with recovery actions, providing feedback loops for error recovery. | 3 / 3 |
Progressive Disclosure | A clear overview points to one-level-deep references (flow-framework.md, bibliography.md, prompts/README.md, and per-stage prompt directories) with explicit load conditions; the on-disk bundle matches the documented counts (find=5, optimize=21, win=3, leverage=1, 30 total), and "Load on demand. Do NOT load all at startup" signals navigation clearly. | 3 / 3 |
Total | 11 / 12 Passed |