CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

lightgbm-analysis

Use when training a LightGBM model on tabular data in R and returning model metrics, feature importance ranking tables, and feature importance plots.

Quality

84%

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

SKILL.md
Quality
Evals
Security

Quality

Content

70%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.

A highly actionable, workflow-sound skill body with explicit validation and error-recovery contracts, undermined by token bloat (field listings and seven examples duplicating cli-guide.md) and a navigation defect: the bundled tests/data/ sample files referenced throughout — including the recommended validation smoke test — are missing from the bundle.

Suggestions

Trim the body by moving the output field listings (importance, metrics, remediation fields) and the Modeling Arguments table into references/cli-guide.md, keeping only the Core Arguments table inline.

Reduce the seven example invocations to two or three representative ones (smoke test, full binary, regression) and point to references/cli-guide.md for the rest.

Either add the referenced tests/data/ sample files (dt_sample1.csv, dt_sample2.csv, dt_sample3.txt) to the bundle or remove the examples and file-table entries that depend on them, since the recommended validation smoke test currently cannot run.

DimensionReasoningScore

Conciseness

The 356-line body is dense but carries substantial trimmable bulk: ~35 lines listing every output field name (importance, metrics, and remediation table fields), seven near-duplicate example invocations where two or three would cover the cases, and inline sample CSV rows — much of it duplicating references/cli-guide.md. It sits between anchor 2 (several unnecessary sections) and anchor 3 (mostly efficient, could be tightened): the prose itself is not padded fluff, but roughly a quarter of the body belongs in reference files.

3 / 5

Actionability

Commands are concrete and complete: a primary invocation with placeholders, two full argument tables, explicit error codes, and seven copy-paste-ready Rscript examples. It falls short of anchor 5 because the examples depend on bundled data under tests/data/ that does not exist in the bundle, so the designated smoke-test validation path is not actually runnable as written.

4 / 5

Workflow Clarity

The "Minimal Workflow" gives a numbered seven-step sequence with pre-run checks (identifier removal, overwrite protection), the "Success And Failure Contract" defines exact terminal signals ("Console output should end with LightGBM analysis completed successfully", required output files, non-zero importance values), and "Caution Remediation" plus rerun_hint provide explicit validate→diagnose→retry feedback loops for each failure code. This matches the top anchor: clear sequence, explicit validation steps, and error-recovery loops.

5 / 5

Progressive Disclosure

References exist, are one level deep, and are clearly signaled via the "Read These Files When Needed" table — but the body inlines substantial content that duplicates references/cli-guide.md (both argument tables, field listings, and examples), and the file table plus multiple examples point to tests/data/ sample files that are absent from the bundle, leaving navigation partially broken. This lands between anchor 3 (content that should be separate is inline, organization could improve) and anchor 4 (mostly appropriate placement, minor gaps): the duplication and broken tests/data path push it to 3.

3 / 5

Total

15

/

20

Passed

Description

92%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.

A well-formed description with an explicit "Use when" trigger, concrete capability list, and a distinct LightGBM/R/tabular niche. The only weakness is trigger-term coverage: no synonyms like "gradient boosting", "GBM", "classification/regression", or data-file extensions.

DimensionReasoningScore

Specificity

The description names four concrete actions — "training a LightGBM model on tabular data in R", "returning model metrics, feature importance ranking tables, and feature importance plots" — which comprehensively covers the skill's purpose. It matches the top anchor (multiple specific concrete actions, comprehensive) rather than anchor 4, which would require notable coverage gaps.

5 / 5

Completeness

It explicitly answers both questions in one sentence: "when" via the explicit trigger "Use when training a LightGBM model on tabular data in R" and "what" via "returning model metrics, feature importance ranking tables, and feature importance plots". This mirrors the top anchor's structure (concrete trigger phrases plus a full capability list); anchor 4 would require the 'when' to be less explicit or specific.

5 / 5

Trigger Term Quality

Good natural keywords: "LightGBM", "train", "model", "tabular data", "R", "feature importance" — phrases a user with this need would actually say. It falls short of anchor 5 because common synonyms and variations are missing, e.g. "gradient boosting", "boosting", "classification", "regression", or file extensions like ".csv"/".tsv".

4 / 5

Distinctiveness Conflict Risk

The named library (LightGBM), language (R), and data type (tabular) define a clear niche with minimal conflict risk — a user wanting SHAP analysis or a Python model would not naturally trigger it, paralleling the PDF-skills anchor example. Anchor 4 ('minor overlap risk with closely related skills') would apply if the trigger were broad enough to collide with sibling ML skills, but the LightGBM qualifier prevents that.

5 / 5

Total

19

/

20

Passed

Validation

93%

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

Validation — 15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

frontmatter_unknown_keys

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

Warning

Total

15

/

16

Passed

Repository

aipoch/medical-research-skills

Path

awesome-med-research-skills/Data Analysis/LightGBM-analysis/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.