detecting-data-anomalies
Process identify anomalies and outliers in datasets using machine learning algorithms. Use when analyzing data for unusual patterns, outliers, or unexpected deviations from normal behavior. Trigger with phrases like "detect anomalies", "find outliers", or "identify unusual patterns".
59
70%
Does it follow best practices?
Impact
—
No eval scenarios have been run
Passed
No known issues
Optimize this skill with Tessl
npx tessl skill review --optimize ./plugins/ai-ml/anomaly-detection-system/skills/detecting-data-anomalies/SKILL.mdQuality
Discovery
89%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 is a solid description with strong trigger terms and explicit 'Use when' and 'Trigger with' clauses that make it highly functional for skill selection. Its main weakness is the lack of specificity in concrete actions—it could benefit from listing specific capabilities like statistical methods used, supported data formats, or output types. The opening word 'Process' seems grammatically awkward and may be a fragment.
Suggestions
Add more specific concrete actions, e.g., 'Applies isolation forests, z-score analysis, and DBSCAN clustering to detect anomalies; generates anomaly reports with severity scores and visualizations.'
Fix the grammatically awkward opening ('Process identify...') to something clearer like 'Identifies anomalies and outliers in datasets...'
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | The description names the domain (anomaly detection in datasets) and mentions 'machine learning algorithms' but doesn't list specific concrete actions beyond 'identify anomalies and outliers.' It lacks details about what specific ML algorithms, what outputs are produced, or what data formats are supported. | 2 / 3 |
Completeness | Clearly answers both 'what' (identify anomalies and outliers in datasets using ML algorithms) and 'when' (explicit 'Use when' clause and 'Trigger with' clause specifying exact phrases and scenarios). | 3 / 3 |
Trigger Term Quality | Good coverage of natural trigger terms: 'detect anomalies', 'find outliers', 'identify unusual patterns', 'unusual patterns', 'outliers', 'unexpected deviations from normal behavior'. These are terms users would naturally use when seeking this capability. | 3 / 3 |
Distinctiveness Conflict Risk | The description carves out a clear niche around anomaly/outlier detection specifically, with distinct trigger terms that are unlikely to conflict with general data analysis or other ML skills. | 3 / 3 |
Total | 11 / 12 Passed |
Implementation
50%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill provides a well-organized overview of anomaly detection with good algorithm selection guidance and a useful error handling table, but falls short on actionability due to the complete absence of executable code examples. The workflow is logically sequenced but lacks validation checkpoints and feedback loops critical for an iterative ML process. Referenced bundle files don't exist, undermining the progressive disclosure structure.
Suggestions
Add executable Python code snippets for at least one algorithm (e.g., Isolation Forest end-to-end), including data loading, scaling, fitting, scoring, and result interpretation—this is the single biggest gap.
Insert explicit validation checkpoints in the workflow, such as 'Verify anomaly score distribution before applying threshold' and 'Review top-10 flagged anomalies with domain context before finalizing contamination parameter.'
Create the referenced bundle files (implementation.md, errors.md) or remove the references if they don't exist, and move the detailed error table and resources into those reference files to improve progressive disclosure.
Trim the Prerequisites and Resources sections—Claude doesn't need to be told what pandas is or how pip install works; focus on version constraints and non-obvious dependencies only.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The content is reasonably structured but includes some unnecessary verbosity—the Prerequisites section explains what pip packages do, the Overview restates the description, and the Resources section lists general knowledge Claude already has. The error handling table and examples add value but could be tighter. | 2 / 3 |
Actionability | The instructions provide a clear 10-step process with algorithm selection guidance, but lack any executable code examples. Everything is described in prose rather than demonstrated with copy-paste-ready Python snippets. The examples section describes scenarios but provides no actual code implementations. | 2 / 3 |
Workflow Clarity | The 10-step workflow is clearly sequenced and logically ordered, but lacks explicit validation checkpoints or feedback loops. There's no 'verify model fit quality before proceeding' step, no guidance on what to do if anomaly scores look wrong, and no iterative refinement loop despite this being a process that inherently requires tuning. | 2 / 3 |
Progressive Disclosure | The skill references `${CLAUDE_SKILL_DIR}/references/implementation.md` and `errors.md` for deeper content, which is good structure. However, no bundle files are provided, so these references are broken. The main file also includes substantial inline content (full error table, detailed examples, resources) that could be better distributed across reference files. | 2 / 3 |
Total | 8 / 12 Passed |
Validation
81%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 9 / 11 Passed
Validation for skill structure
| Criteria | Description | Result |
|---|---|---|
allowed_tools_field | 'allowed-tools' contains unusual tool name(s) | Warning |
frontmatter_unknown_keys | Unknown frontmatter key(s) found; consider removing or moving to metadata | Warning |
Total | 9 / 11 Passed | |
- Commit
196527a
- 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.