Content
43%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 admirably brief but underspecified: it relies on vague, sometimes ungrammatical directives instead of executable guidance, presents an unsequenced workflow for what is explicitly an iterative verify-and-proceed process, and lacks section organization. The core behavioral idea (skeptically verify facts before proceeding) is good, but the skill never tells Claude concretely how to do any of it.
Suggestions
Sequence the workflow into numbered phases (e.g., 1. Read research discoveries; 2. Draft plan sections in .cursor/plans/<name>/plan.md; 3. For each unverified fact, spawn a research task; 4. Ask the user only for what research cannot resolve; 5. Proceed) so the iterative loop is explicit.
Make the directives executable: replace "Spawn new research tasks to verify the correct information (research/command)" with the concrete command/task format to use, and specify what the to-do list and plan .md file should contain.
Add section headers (Goal / Output / Process) and fix the garbled sentence "Reading the discoveries, during the research, make a plan" — poor organization and broken grammar directly hurt navigation and actionability.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is short (~18 lines) with no explanation of concepts Claude already knows, so it largely respects the token budget. Minor padding remains — the role-setting opener "You should be skeptical, thorough, and work collaboratively with the user to produce high-quality technical specifications" and the redundant "Reading the discoveries, during the research" phrasing — matching 'Efficient; minor instances of over-explanation that could be trimmed' rather than 5, where every token would earn its place. | 4 / 5 |
Actionability | Concrete guidance is minimal: the only specific instruction is the output location ".cursor/plans/{plan_name_folder} using a .md file style", while directives like "Spawn new research tasks to verify the correct information (research/command)" and "DO NOT just accept the correction or suggestion from user" are high-level, cryptic, or ungrammatical, with no steps showing how to execute them. This matches 'Minimal concrete guidance; high-level hints but missing the specific steps to execute' rather than 1, because it does instruct with at least one concrete artifact path rather than merely describing. | 2 / 5 |
Workflow Clarity | The imperatives exist but are an unsequenced bullet list, and the ordering is incoherent ("Reading the discoveries, during the research, make a plan" gives no clear phase ordering of research vs. planning vs. verification). There are no checkpoints or explicit loop structure beyond "Only proceed once you've verified the facts yourself", matching 'Rough sequence present but many gaps; steps poorly defined; validation absent' rather than 3, because even the step list itself lacks a defined order. | 2 / 5 |
Progressive Disclosure | The skill is under 50 lines with no bundle files, so no external references are needed, but the body itself is not well-organized: it has only the title heading, two loosely connected prose paragraphs (one garbled), and a flat bullet list with no section headers separating goal, output format, and process rules. This matches 'Some structure but could be better organized' rather than 5, which per the simple-skill guideline requires well-organized sections, and rather than 4 because the organization gaps are not minor. | 3 / 5 |
Total | 11 / 20 Passed |