outlier-detection-and-quality-assessment

执行全面的异常值检测与数据质量评估，利用 IQR 方法识别异常值并结合偏度、峰度分析数据分布特征，适用于非正态分布数据的预处理阶段。

Quality

70%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./skills/sn-da-excel-workflow/capability/excel-data-cleaning/outlier-detection/SKILL.md

Quality

Content

80%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The skill delivers executable, concise, copy-paste-ready code across a clear 5-step sequence, but the destructive outlier-handling step lacks validation checkpoints and the ~150-line body is monolithic with no progressive disclosure. Adding verification and optionally splitting advanced material would lift the weaker dimensions.

Suggestions

Add a validation checkpoint after Step 5's cap/drop handling — e.g. print and compare df.shape / distribution stats before and after, and confirm no unintended data loss before proceeding.

Move the Step 3 visualization and Step 5 handling skeleton into a referenced file (e.g. references/handling.md) so SKILL.md stays a concise overview with one-level-deep links.

Surface the file_path placeholder and key outputs (outlier_analysis_report.png, analysis_df) up front so the workflow's inputs and artifacts are unambiguous.

Dimension	Reasoning	Score
Conciseness	The body is code-forward and lean: brief step headers plus executable Python, with only short navigational comments — it does not explain concepts Claude already knows (no 'what is IQR' or 'what is skewness').	3 / 3
Actionability	Steps 1–4 are complete, copy-paste-ready executable Python with a clearly marked file_path placeholder, and Step 5 provides a runnable handle_outliers() function with example usage.	3 / 3
Workflow Clarity	Five steps are clearly sequenced (Step 1 → Step 5), but the data-modifying Step 5 (cap/drop outliers) is a destructive/batch operation with no validation checkpoint (e.g. before/after row counts or distribution check), which caps workflow clarity at 2.	2 / 3
Progressive Disclosure	Content is organized into five clear sections, but at ~150 lines it is a single monolithic file with no external references or progressive disclosure, exceeding the 'under 50 lines' simple-skill exception that would otherwise allow a 3.	2 / 3
	Total	10 / 12 Passed

Description

60%

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description is specific about capabilities and uses the right domain terms, but lacks an explicit 'Use when…' trigger and leans on technical jargon over natural user phrasing. It is distinguishable but not maximally distinct from generic data-cleaning skills.

Suggestions

Add an explicit trigger clause, e.g. 'Use when the user needs to detect outliers, assess data quality, or clean skewed/non-normal datasets before modeling.'

Soften jargon by including natural variations users would actually say (e.g. 'find outliers in my data', 'check data distribution').

Tighten distinctiveness by leading with the core action ('Detect outliers using the IQR method') before the broader 'data quality assessment' framing.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — '异常值检测', '数据质量评估', '利用 IQR 方法识别异常值', '偏度、峰度分析数据分布特征' — comparable to the multi-action 'Extract text and tables, fill forms, merge documents' anchor.	3 / 3
Completeness	Clearly states 'what' (outlier detection via IQR + skewness/kurtosis quality assessment) but only implies 'when' via '适用于非正态分布数据的预处理阶段' — no explicit 'Use when…' trigger clause, which the guidelines cap at 2.	2 / 3
Trigger Term Quality	Contains relevant terms a data analyst would say ('异常值', 'IQR', '偏度', '峰度', '数据质量') but is jargon-dense ('非正态分布数据的预处理阶段') and misses common natural variations like 'find outliers in my data' or 'clean my dataset'.	2 / 3
Distinctiveness Conflict Risk	The IQR + skewness/kurtosis niche is fairly specific, but '数据质量评估' and the preprocessing framing could overlap with general EDA/data-cleaning skills, so it is not yet a clear distinct trigger.	2 / 3
	Total	9 / 12 Passed

Validation

100%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: OpenSenseNova/SenseNova-Skills
Path: skills/sn-da-excel-workflow/capability/excel-data-cleaning/outlier-detection/SKILL.md
Commit: 179fea1

Reviewed: 2 days ago

Table of Contents

Discovery Implementation Validation

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.