Content
32%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 a project specification rather than an operational skill: it describes capabilities, implementation status, and roadmap, but gives the model no sequenced workflow and no supporting reference files. Concrete details (CLI examples, thresholds, error codes) are diluted by large amounts of project-management padding. It needs restructuring into a lean procedural core with detail moved to reference files.
Suggestions
Replace the descriptive sections (Implementation Status, Contributing, Research Sources) with an ordered extraction workflow: detect stack -> extract tokens -> extract components -> generate outputs -> run quality gates, with explicit validation checkpoints.
Move the output-directory spec, multi-AI orchestration rules, and algorithm details into references/ files (e.g. references/output-structure.md, references/multi-ai.md) and keep SKILL.md as a lean overview with one-level-deep, clearly signaled links.
Trim the capability bullet lists to what the model must actually do, and add the concrete commands/scripts for each pipeline stage so the guidance is executable rather than descriptive.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Several sections are padded relative to what a model needs to execute the skill: marketing-style capability lists ('Pattern Detection: Layout patterns...'), 'Implementation Status' checklists, a 'Contributing' section with an 8-week roadmap, and external 'Research Sources' links. These are project-management artifacts, not skill instructions, matching 'noticeably verbose; several unnecessary explanations or padded sections' rather than the mostly-efficient anchor at 3. | 2 / 5 |
Actionability | There is genuinely concrete material (CLI invocations like '/octo:extract ./my-app --mode design', the output directory tree, error codes ERR-001..VAL-004, quality-gate thresholds, and algorithm parameters like 'CIEDE2000', 'k=8 clusters', 'ΔE < 2'). But the guidance mostly describes a system rather than instructing how to run it — the actual pipeline steps, tooling, and commands for executing an extraction are absent (the skill admits it is a 'Skeleton'), fitting 'some concrete guidance but incomplete; missing key details'. | 3 / 5 |
Workflow Clarity | There is no sequenced workflow at all: no ordered steps from input to validated output, no checkpoints, and the quality gates are listed as isolated rules rather than wired into a process. The 'Usage Patterns' section shows invocations but not what happens next. This fits 'rough sequence present but many gaps' more than the steps-listed-with-validation-gaps anchor at 3, since even the steps themselves are missing. | 2 / 5 |
Progressive Disclosure | No bundle files exist (references/, scripts/, assets/ are all absent), so everything — the full output-structure spec, algorithm details, multi-AI orchestration rules, roadmap, and research links — is inlined in one ~230-line SKILL.md. Content that clearly belongs in separate reference files (e.g. the output layout spec and provider orchestration rules) is inlined, matching the anchor at 2; it avoids a 1 only because section headers make the monolith navigable. | 2 / 5 |
Total | 9 / 20 Passed |