Content
78%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 content is a well-structured, actionable instruction skill: concrete openspec commands, a clear numbered workflow with a stop checkpoint, and a tight Guardrails section with no padded explanation. It misses the top marks mainly on conciseness and a few actionability edges, and workflow_clarity could reach 5 with an explicit validation/retry loop.
Suggestions
Tighten emphasis and restatement: replace "**IMPORTANT**: Do NOT proceed without understanding what the user wants to build" with a single concise directive, and merge the repeated kebab-case input rules into one place.
Add a lightweight validation/retry checkpoint after `openspec new change` (e.g., confirm the directory was created and surface errors before proceeding to `openspec status`) to push workflow_clarity to 5.
Make the command examples fully copy-paste ready by clarifying how `<name>` and `<first-artifact-id>` are substituted from prior command output rather than leaving them as placeholders.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body avoids explaining concepts Claude already knows and keeps each step tight and purposeful; minor emphasis tokens ("**IMPORTANT**", "**Otherwise**") and slight restatement of the input rule could still be trimmed, so it sits just below lean/efficient. | 4 / 5 |
Actionability | It provides concrete, runnable commands (openspec new change, status, instructions, schemas --json) plus a worked name-derivation example, with only minor gaps such as placeholder substitution guidance; fully actionable but not copy-paste-complete across all branches. | 4 / 5 |
Workflow Clarity | A clear 6-step sequence with an explicit STOP checkpoint and guardrails for invalid names and existing changes; it lacks an explicit validate-then-retry feedback loop, but since the operation is non-destructive scaffolding the destructive-cap does not apply, keeping it at 4 rather than 5. | 4 / 5 |
Progressive Disclosure | The skill is under 50 lines with no need for external references and is cleanly organized into Input, Steps, Output, and Guardrails sections, meeting the simple-skill exception for a top progressive-disclosure score. | 5 / 5 |
Total | 17 / 20 Passed |