gtars
High-performance toolkit for genomic interval analysis in Rust with Python bindings. Use when working with genomic regions, BED files, coverage tracks, overlap detection, tokenization for ML models, or fragment analysis in computational genomics and machine learning applications.
55
62%
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 ./scientific-skills/gtars/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 strong description with excellent trigger term coverage and completeness, clearly targeting a specialized computational genomics niche. Its main weakness is that the capabilities are described more as topic areas than concrete actions — listing specific operations (e.g., 'merge intervals', 'compute coverage depth', 'convert BED to tokenized sequences') would improve specificity. Overall it would perform well in skill selection due to its distinctive domain and explicit 'Use when' clause.
Suggestions
Replace or supplement the topic-area phrases with concrete action verbs, e.g., 'Computes coverage depth, merges overlapping intervals, tokenizes genomic regions for ML models, detects interval overlaps' instead of 'overlap detection, tokenization for ML models'.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Names the domain (genomic interval analysis) and mentions several areas like 'overlap detection', 'tokenization for ML models', 'fragment analysis', but these read more as topic areas than concrete actions. It doesn't list specific operations like 'compute coverage from BAM files' or 'merge overlapping intervals'. | 2 / 3 |
Completeness | Clearly answers both 'what' (high-performance toolkit for genomic interval analysis in Rust with Python bindings) and 'when' (explicit 'Use when...' clause listing specific trigger scenarios like BED files, coverage tracks, overlap detection, tokenization for ML, fragment analysis). | 3 / 3 |
Trigger Term Quality | Includes strong natural keywords users would actually use: 'genomic regions', 'BED files', 'coverage tracks', 'overlap detection', 'tokenization', 'ML models', 'fragment analysis', 'computational genomics', 'Rust', 'Python bindings'. These cover a wide range of terms a user in this domain would naturally mention. | 3 / 3 |
Distinctiveness Conflict Risk | Highly distinctive niche — genomic interval analysis with Rust/Python bindings is very specific and unlikely to conflict with other skills. The combination of genomics-specific terms (BED files, coverage tracks, genomic regions) and implementation details (Rust, Python bindings) creates a clear, unique identity. | 3 / 3 |
Total | 11 / 12 Passed |
Implementation
35%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill is well-structured as a table of contents for a complex toolkit, with clear module organization and references to deeper documentation. However, it is significantly too verbose, with many sections explaining concepts Claude already knows and 'when to use' lists that add little value. The code examples appear plausible but potentially inaccurate, and workflows lack validation checkpoints that would be important for file-processing operations.
Suggestions
Cut the 'When to use' bullet lists, 'Performance Characteristics', 'Data Formats', and 'Python vs CLI Usage' sections entirely — Claude knows these concepts and they consume tokens without adding actionable guidance.
Verify all Python API examples against the actual gtars API (e.g., confirm RegionSet.from_bed, igd.build_index, filter_overlapping exist) and replace any pseudocode with tested, executable snippets.
Add validation/verification steps to workflows — e.g., after generating a coverage track, check file size or load a region to confirm correctness; after tokenization, verify token count matches expectations.
Either provide the referenced bundle files (references/overlap.md, etc.) or remove the references and inline the most critical information to avoid dead links.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is highly verbose at ~200+ lines. It includes unnecessary explanations Claude already knows (what BED files are, when to use coverage tracks, what 'native Rust performance' means). The 'Performance Characteristics', 'Data Formats', 'Python vs CLI Usage', and 'When to use' bullet lists are padding that doesn't add actionable value. The 'Use this skill when working with' section in the overview repeats what the module sections already cover. | 1 / 3 |
Actionability | Code examples are provided for each module, but many appear to be pseudocode or aspirational rather than verified executable code. For instance, `gtars.RegionSet.from_bed()`, `peaks.filter_overlapping()`, and `gtars.igd.build_index()` may not reflect the actual API. The CLI examples look plausible but lack verification. The ML preprocessing workflow's Step 4 is a comment placeholder rather than concrete guidance. | 2 / 3 |
Workflow Clarity | Three workflows are listed with numbered steps, which is good. However, none include validation checkpoints or error recovery steps. The coverage track pipeline has two steps but no verification that output is correct. The ML preprocessing workflow ends with a vague comment. No feedback loops exist for any workflow despite dealing with file transformations. | 2 / 3 |
Progressive Disclosure | The skill references six separate reference files (references/overlap.md, references/coverage.md, etc.) which is good structure, but no bundle files are provided, so these references are unverifiable dead links. The main file itself contains too much inline content that could be in reference files (Performance Characteristics, Data Formats, Error Handling sections), while the overview sections for each module are somewhat redundant with the referenced docs. | 2 / 3 |
Total | 7 / 12 Passed |
Validation
90%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 10 / 11 Passed
Validation for skill structure
| Criteria | Description | Result |
|---|---|---|
metadata_version | 'metadata.version' is missing | Warning |
Total | 10 / 11 Passed | |
- Repository
- K-Dense-AI/claude-scientific-skills
- Commit
cbcae7b
- 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.