CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

data-transform

Transform, clean, reshape, and preprocess data using pandas and numpy. Works with ANY LLM provider (GPT, Gemini, Claude, etc.).

Quality

62%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./scientific-skills/Data Analysis/data-transform/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

53%Weight 40%Scale 1-5

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

The body is highly actionable — abundant, executable pandas examples with a sensible step order — but it buries a small amount of domain-specific material (AnnData, CPM, DEG processing) inside a long generic pandas tutorial Claude does not need. It should be restructured as a lean overview with the cookbook detail moved to reference files, with validation checkpoints added for batch operations.

Suggestions

Move the standard-pandas cookbook (cleaning, scaling, reshaping, filtering, merging, type conversions) into references/ files and keep SKILL.md as a concise overview linking to them — the current monolith scores the context window heavily for content Claude already knows.

Add explicit validation checkpoints, especially in "Batch Processing Multiple Files" (e.g. verify row counts/schema per file and handle read errors) to satisfy the batch-operation validation requirement.

Delete the "Critical API Reference - DataFrame vs Series Attributes" section and the "✅" print padding — they explain pandas basics Claude already knows and add tokens without adding guidance.

Fix or flag deprecated patterns shown in examples: df.fillna(method='ffill') and fillna(..., inplace=True) on a column slice warn or fail on pandas 2.x.

DimensionReasoningScore

Conciseness

The ~575-line body is largely a pandas cookbook of operations Claude already knows (groupby/agg, merge/concat, melt/pivot, pd.cut, get_dummies, astype, string ops) plus a tutorial section on ".dtype vs .dtypes - Common Pitfall!" and decorative "print('✅ ...')" padding. Only the AnnData conversion, CPM normalization, and DEG-processing sections add non-obvious value, matching anchor 2 (noticeably verbose, several unnecessary explanations or padded sections).

2 / 5

Actionability

Code throughout is concrete, complete, and executable — e.g. the remove_outliers IQR function, "merged = pd.merge(df1, df2, left_on='gene_name', right_on='gene_id')", and the batch glob loop — with comments covering the common transformation cases. Fits anchor 5 (copy-paste ready, covers common cases).

5 / 5

Workflow Clarity

Steps 1-6 give a real sequence (import → clean → normalize → reshape → filter → merge) and the Troubleshooting section offers recovery patterns, but validation is only implicit ("Check Data First" sits in best practices, not as a checkpoint step). The "Batch Processing Multiple Files" loop has no error handling or validation, which caps workflow clarity at 3 per the batch-operations rule.

3 / 5

Progressive Disclosure

The body is monolithic: 400+ lines of inlined API cookbook that clearly belongs in separate reference files, no bundle files exist, and the References section points only to external documentation URLs. Matches anchor 2 (content that clearly belongs in separate files is inlined); the section headers keep it above 1.

2 / 5

Total

12

/

20

Passed

Description

55%Weight 40%Scale 1-5

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 communicates a clear set of capabilities and the libraries used, but it lacks any "Use when..." trigger guidance and misses common natural synonyms for data-preparation tasks. The appended claim about working with any LLM provider adds noise without describing a user-facing capability.

Suggestions

Add an explicit trigger clause, e.g. "Use when the user asks to clean, normalize, merge, filter, or reshape tabular data, or mentions pandas, missing values, or CSV files."

Remove or shorten the "Works with ANY LLM provider (GPT, Gemini, Claude, etc.)" sentence — it is a marketing claim, not a capability, and its provider keywords risk triggering for unrelated LLM-configuration tasks.

Include natural synonyms users actually say (normalize, scale, missing values, merge, filter, pivot, wide/long format) to improve keyword coverage.

DimensionReasoningScore

Specificity

"Transform, clean, reshape, and preprocess data using pandas and numpy" lists several concrete action verbs and names the tools, matching anchor 4 (several specific actions, minor gaps). Not 5 because it omits many covered capabilities (merge, filter, missing values, feature engineering); not 3 because it names more than 1-2 actions plus the implementation tools.

4 / 5

Completeness

The "what" is clear (transform/clean/reshape/preprocess data via pandas and numpy), but there is no "Use when..." or equivalent explicit trigger clause, which the judging guidelines cap at 3. Not 4 because the when to use it is entirely absent, not just weakly implied.

3 / 5

Trigger Term Quality

"clean", "transform", "preprocess", and "pandas" are natural user phrases, but common variations are missing ("normalize", "missing values", "merge datasets", ".csv"), and the trailing "Works with ANY LLM provider (GPT, Gemini, Claude, etc.)" contributes irrelevant keywords rather than triggers. Fits anchor 3 (some relevant keywords, missing common variations).

3 / 5

Distinctiveness Conflict Risk

"Transform, clean, reshape, and preprocess data" is a broad data-work description that could overlap with analysis, EDA, or plotting skills; the pandas/numpy naming provides some anchoring, matching anchor 3 (somewhat specific but could still overlap with similar skills).

3 / 5

Total

13

/

20

Passed

Validation

87%

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

Validation — 14 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

skill_md_line_count

SKILL.md is long (589 lines); consider splitting into references/ and linking

Warning

frontmatter_unknown_keys

Unknown frontmatter key(s) found; consider removing or moving to metadata

Warning

Total

14

/

16

Passed

Repository

aipoch/medical-research-skills

Path

scientific-skills/Data Analysis/data-transform/SKILL.md

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.