review-and-refactor
Review and refactor code in your project according to defined instructions
41
11%
Does it follow best practices?
Impact
96%
0.98xAverage score across 3 eval scenarios
Passed
No known issues
Optimize this skill with Tessl
npx tessl skill review --optimize ./skills/review-and-refactor/SKILL.mdQuality
Discovery
14%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This description is too vague and generic to effectively differentiate itself from other code-related skills. It lacks specific concrete actions, natural trigger term variations, and an explicit 'Use when...' clause. The phrase 'according to defined instructions' adds no useful information for skill selection.
Suggestions
Add a 'Use when...' clause with specific trigger scenarios, e.g., 'Use when the user asks to clean up code, reduce technical debt, improve code structure, or perform a code review.'
List specific concrete actions the skill performs, e.g., 'Identifies code smells, extracts reusable functions, renames variables for clarity, simplifies complex logic, and removes dead code.'
Clarify what 'defined instructions' means — specify whether these are project-level style guides, linting rules, or user-provided refactoring criteria to make the skill more distinctive.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | The description mentions 'review and refactor code' which are two actions, but they are very broad and generic. 'According to defined instructions' is vague and doesn't specify what kind of refactoring or reviewing is performed. No concrete actions like 'rename variables', 'extract functions', 'identify code smells', etc. | 1 / 3 |
Completeness | The description weakly addresses 'what' (review and refactor code) but has no explicit 'when' clause or trigger guidance. There is no 'Use when...' statement, which per the rubric should cap completeness at 2, and the 'what' is itself quite vague, bringing this to a 1. | 1 / 3 |
Trigger Term Quality | 'Review' and 'refactor' are natural terms users might say, but the description lacks common variations like 'clean up code', 'improve code quality', 'code smell', 'technical debt', 'restructure', or 'code review'. The phrase 'defined instructions' is not a natural user term. | 2 / 3 |
Distinctiveness Conflict Risk | 'Review and refactor code' is extremely generic and would overlap with virtually any code-related skill — linting skills, code quality skills, testing skills, general coding assistance skills. 'According to defined instructions' doesn't narrow the scope meaningfully. | 1 / 3 |
Total | 5 / 12 Passed |
Implementation
7%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill is extremely vague and provides almost no actionable guidance. It essentially says 'read the instructions and refactor the code' without specifying what kinds of refactorings to look for, how to prioritize changes, or how to validate results. The 'take a deep breath' prompt engineering and role-playing preamble waste tokens without adding value.
Suggestions
Replace the vague 'review and refactor' instruction with specific refactoring patterns to look for (e.g., dead code removal, naming conventions, duplication) or a concrete checklist of what to evaluate.
Add a validation workflow: e.g., run tests before changes, make changes, run tests after, diff the changes, and verify no behavioral regressions.
Remove the 'take a deep breath' phrase and the 'Role' section — these waste tokens and don't improve Claude's performance.
Add concrete examples of what a refactoring pass looks like: e.g., before/after code snippets or a sample review comment format.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The 'take a deep breath' prompt engineering phrase is unnecessary for Claude. The 'Role' section telling Claude it's a 'senior expert software engineer' is padding. The instructions are vague and padded with filler rather than being lean and efficient. | 1 / 3 |
Actionability | The skill provides only vague, abstract direction ('review all the code carefully and make code refactorings if needed') with no concrete commands, code examples, or specific refactoring patterns. It describes rather than instructs. | 1 / 3 |
Workflow Clarity | The numbered steps are superficial and lack any validation checkpoints. There's no guidance on how to verify refactorings are correct, no feedback loop for test failures, and no clear sequence for handling conflicts between different instruction files. | 1 / 3 |
Progressive Disclosure | The skill references external instruction files (`.github/instructions/*.md` and `.github/copilot-instructions.md`) which is appropriate delegation, but there's no description of what those files contain or how to navigate them. The skill itself is short enough to not need splitting, but the references are not well-signaled. | 2 / 3 |
Total | 5 / 12 Passed |
Validation
100%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 11 / 11 Passed
Validation for skill structure
No warnings or errors.
- Repository
- github/awesome-copilot
- Commit
2e7c43b
- Reviewed
Table of Contents
Is this your skill?
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.