CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

outlier-detection-handler

Use outlier detection handler for data analysis workflows that need structured execution, explicit assumptions, and clear output boundaries.

Quality

54%

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/outlier-detection-handler/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

42%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 skill body is drowning in generic policy boilerplate that duplicates itself across three overlapping output/failure sections, while the actual domain content (how to run the outlier detector, what the methods do, a real invocation) is thin. Structure exists, but the token budget is spent on templated governance text rather than executable guidance for the one bundled script.

Suggestions

Add a copy-paste-ready example invocation of the core task, e.g., 'python scripts/main.py data/patients.csv --method IQR --action winsorize', with a sample of its output.

Consolidate the redundant 'Error Handling'/'Failure Handling' and 'Output Requirements'/'Response Template'/'Output Contract' sections into one error-and-output section, and trim generic boilerplate (Security Checklist, Evaluation Criteria, Lifecycle Status) that applies to any skill.

Keep SKILL.md to a concise overview (usage, parameters, example, fallback) and move the detailed policy/checklist material into a one-level-deep reference file if it must be retained.

DimensionReasoningScore

Conciseness

The ~190-line body contains numerous padded, redundant policy sections: 'Error Handling' and 'Failure Handling' repeat the same guidance, 'Output Requirements', 'Response Template', and 'Output Contract' overlap heavily, and generic boilerplate like 'Security Checklist', 'Evaluation Criteria', and 'Lifecycle Status' adds little skill-specific value. This matches the score-2 anchor ('noticeably verbose; several unnecessary explanations or padded sections') and is well below the efficient score-4/5 anchors.

2 / 5

Actionability

There is some concrete guidance — the py_compile/--help commands, a parameter table, and a brief worked example ('Input: Biomarker measurements from 200 patients / Output: 5 outliers identified (2.5%)...') — but no executable example of the core task: an actual invocation like 'python scripts/main.py data.csv --method IQR --action winsorize' is missing. Guidance is partially concrete but incomplete, matching the score-3 anchor and falling short of score 4's 'mostly executable guidance'.

3 / 5

Workflow Clarity

The Workflow section lists a coherent 5-step sequence with fallback handling ('If execution fails or inputs are incomplete, switch to the fallback path'), and the Quick Check provides a validation command. However, the steps themselves are abstract ('Confirm the user objective... validate that the request matches the documented scope') rather than executable checkpoints, and the data-modifying actions (remove/winsorize, file writes) lack an explicit validate-then-proceed loop, matching the score-3 anchor ('steps listed but validation gaps; checkpoints missing or implicit').

3 / 5

Progressive Disclosure

The body has clear section headers and its script references are real (scripts/main.py exists), but everything — security checklists, risk tables, evaluation criteria, required-inputs tables — is inlined in a single ~190-line SKILL.md with no one-level-deep reference files for the bulk material. This matches the score-3 anchor ('some structure but could be better organized; content that should be separate is inline') and falls short of score 4's 'most content is appropriately placed'.

3 / 5

Total

11

/

20

Passed

Description

45%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 leans on generic workflow boilerplate instead of stating what the skill concretely does. It never mentions detecting, flagging, removing, or winsorizing outliers, nor the supported methods or file formats, and the trigger phrase is abstract property language rather than natural user intent. A rewrite stating the concrete capabilities plus an explicit 'Use when...' clause would lift every dimension.

Suggestions

State the concrete capabilities explicitly, e.g., 'Identify statistical outliers in CSV/Excel datasets using 3-sigma, IQR, or Grubbs methods, and flag, remove, or winsorize them.'

Add an explicit trigger clause naming natural user phrases, e.g., 'Use when the user mentions outliers, anomalies, data cleaning, or data quality screening.'

Replace the abstract meta-properties ('structured execution, explicit assumptions, clear output boundaries') with user-facing outcomes such as audit-ready outlier reports for data quality control.

DimensionReasoningScore

Specificity

The description names the domain ('outlier detection handler') but its only capability statements are generic meta-properties ('structured execution, explicit assumptions, and clear output boundaries') rather than concrete actions like detecting, flagging, or winsorizing outliers. It names the domain but the actions are minimal or generic, matching the score-2 anchor and falling clearly short of the score-3 anchor's '1-2 concrete actions'.

2 / 5

Completeness

The 'when' is present but weak ('for data analysis workflows that need structured execution...') — it describes workflow properties rather than concrete situations, and the 'what' is only implied by the skill name since no actual capability is stated. This sits between the score-2 anchor (vague what, no when) and score-4 anchor (both present, when could be more specific), but noticeably below the midpoint because the what-portion states no real action.

3 / 5

Trigger Term Quality

'outlier detection' and 'data analysis' are relevant keywords a user might say, but common variations and synonyms are missing entirely (e.g., 'outliers', 'anomalies', 'data cleaning', file types like CSV/Excel). This matches the score-3 anchor ('some relevant keywords but missing common variations or synonyms') and is below score 4, which requires good coverage with only a few terms missing.

3 / 5

Distinctiveness Conflict Risk

'outlier detection handler' carves out a somewhat specific niche, but the trigger phrasing 'data analysis workflows that need structured execution' is broad and could overlap with many general data-analysis and quality-control skills. This matches the score-3 anchor ('somewhat specific but could still overlap with similar skills') and does not reach score 4, which requires mostly distinct triggers.

3 / 5

Total

11

/

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

scientific-skills/Data Analysis/outlier-detection-handler/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.