Content
100%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
An exemplar instruction-only skill: lean, highly actionable with concrete tool parameters, a clearly sequenced recall workflow with explicit error-handling checkpoints, and well-organized sections appropriate for a self-contained single-purpose skill with no bundle files.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and assumes Claude's competence — it names tools and parameters directly without explaining what memory, wikilinks, or BM25 are, and every section earns its place. | 3 / 3 |
Actionability | Concrete, copy-paste-ready guidance throughout: specific tool names (search, traverse, daily_list, frontmatter_read, read) with exact parameters (query, limit=5, path, depth=1, direction=both, date) and the server start command and connection URL. | 3 / 3 |
Workflow Clarity | The recall workflow is clearly sequenced (pick one of three independent modes, call the matching tool, read hits by path, cite paths) with explicit feedback checkpoints: "If nothing useful comes back, say so plainly" and the missing-tools error-recovery path. | 3 / 3 |
Progressive Disclosure | A self-contained, under-50-line skill with no need for external references, organized into well-signaled sections (Recall, Server status, Workspace model), which meets the simple-skill bar for progressive disclosure. | 3 / 3 |
Total | 12 / 12 Passed |