Data Analysis
lasso-logistics-analysis
Use when building a binary classification model from an expression matrix or other omics feature matrix with LASSO logistic regression, cross-validation, and coefficient path visualization. NOT for: multiclass classification, survival/Cox models, or ordinary linear regression.
85100Total Score
Core Capability
94 / 100
Functional Suitability
11 / 12
Reliability
12 / 12
Performance & Context
8 / 8
Agent Usability
14 / 16
Human Usability
8 / 8
Security
12 / 12
Maintainability
11 / 12
Agent-Specific
18 / 20
Medical Task
21 / 25 Passed
78Binary classification on expression_matrix.csv vs groups.csv
4/5
78Custom --nfolds 5 and --seed 123
4/5
78Optional feature panel restriction from genes.csv
4/5
78Custom CV plot title via --cv_title flag
4/5
81Missing --case_group argument validation test
5/5
Veto GatesRequired pass for any deployment consideration
Skill Veto✓ All 4 gates passed
✓
Operational Stability
System remains stable across varied inputs and edge cases
PASS✓
Structural Consistency
Output structure conforms to expected skill contract format
PASS✓
Result Determinism
Equivalent inputs produce semantically equivalent outputs
PASS✓
System Security
No prompt injection, data leakage, or unsafe tool use detected
PASSResearch Veto✅ PASS — Applicable
| Dimension | Result | Detail |
|---|---|---|
| Scientific Integrity | PASS | No fabricated statistics; all outputs computed from actual input data |
| Practice Boundaries | PASS | NOT for section in frontmatter description prevents misuse for multiclass, survival, and linear regression |
| Methodological Ground | PASS | LASSO binomial glmnet with alpha=1 correctly documented; lambda.min extraction is methodologically correct; no principled fallacies |
| Code Usability | PASS | Dependency check fires cleanly with structured error; 8-file modular structure verified; each file under 150 lines; no syntax errors detected |
Core Capability94 / 100 — 8 Categories
Functional Suitability
Binary classification, LASSO (alpha=1), cv.glmnet, coefficient path and CV plots covered; no alpha control (fixed at 1) limits scope compared to elastic-net-feature-selection
11 / 12
92%
Reliability
15 SKILL_* codes — most comprehensive error coverage in the batch; includes SKILL_PARSE_ERROR, SKILL_FILE_WRITE_ERROR, SKILL_MEMORY_ERROR beyond standard set
12 / 12
100%
Performance & Context
SKILL.md 275 lines; When to Read External Files table at top; modular scripts under 150 lines each; timeout control
8 / 8
100%
Agent Usability
Usage section clear; workflow steps labeled; When to Read at top; minor gaps: no agent response contract, no dedicated stop conditions section in body
14 / 16
88%
Human Usability
Natural trigger language; 15 error codes handle all edge cases; strict validation correct per Category 3 Override 2
8 / 8
100%
Security
requireNamespace() dependency checks; no eval/exec/system; temp file cleanup in implementation checklist; session_info for reproducibility
12 / 12
100%
Maintainability
8-file modular structure with explicit less-than-150-lines-per-file constraint; clean separation of plotting, modeling, io; no post-run file checklist
11 / 12
92%
Agent-Specific
Precise trigger with NOT for in frontmatter; When to Read External Files at top; clean CSV/PDF outputs; idempotent reruns; minor gap: stop conditions not in SKILL.md body
18 / 20
90%
Core Capability Total94 / 100
Medical TaskExecution Average: 78.6 / 100 — Assertions: 21/25 Passed
78
Canonical
Binary classification on expression_matrix.csv vs groups.csv
4/5 ✓
78
Variant A
Custom --nfolds 5 and --seed 123
4/5 ✓
78
Edge
Optional feature panel restriction from genes.csv
4/5 ✓
78
Variant B
Custom CV plot title via --cv_title flag
4/5 ✓
81
Stress
Missing --case_group argument validation test
5/5 ✓
78
Canonical✅ Pass
Binary classification on expression_matrix.csv vs groups.csv
SKILL_DEPENDENCY_MISSING: glmnet — environment constraint; structured error with install instruction confirmed
Basic 32/40|Specialized 46/60|Total 78/100
✅A1SKILL_DEPENDENCY_MISSING error emitted with package name and install instruction
✅A2Script exits with code 1
✅A3Error message includes install instruction
✅A4Basic usage command matches documented Usage section exactly
❌A5coefficient.csv, selected_features.txt, and PDF plots produced
Pass rate: 4 / 5
78
Variant A✅ Pass
Custom --nfolds 5 and --seed 123
SKILL_DEPENDENCY_MISSING: glmnet — consistent error; nfolds valid values documented
Basic 32/40|Specialized 46/60|Total 78/100
✅A1SKILL_DEPENDENCY_MISSING consistent with Input 1
✅A2--nfolds 5 and --seed 123 accepted without validation error
✅A3Script exits cleanly
✅A4nfolds documented with valid values: 3, 5, 7, 10
❌A5CV output with 5-fold cross-validation produced
Pass rate: 4 / 5
78
Edge✅ Pass
Optional feature panel restriction from genes.csv
SKILL_DEPENDENCY_MISSING: glmnet — feature panel behavior documented in Workflow Step 2; missing_features.txt documented
Basic 32/40|Specialized 46/60|Total 78/100
✅A1SKILL_DEPENDENCY_MISSING (correct)
✅A2--feature parameter accepts file path
✅A3Script exits cleanly
✅A4Feature panel behavior documented including missing_features.txt output
❌A5Feature-restricted output produced with missing features listed
Pass rate: 4 / 5
78
Variant B✅ Pass
Custom CV plot title via --cv_title flag
SKILL_DEPENDENCY_MISSING: glmnet — --cv_title accepted; default empty title behavior documented
Basic 32/40|Specialized 46/60|Total 78/100
✅A1SKILL_DEPENDENCY_MISSING (correct)
✅A2--cv_title accepted as optional string
✅A3Script exits cleanly
✅A4Custom plot title documented (left empty by default unless user provides)
❌A5PDF with custom CV title produced
Pass rate: 4 / 5
81
Stress✅ Pass
Missing --case_group argument validation test
Required argument validation fires before dependency check; SKILL_INVALID_PARAMETER covered; 15 error codes comprehensive
Basic 34/40|Specialized 47/60|Total 81/100
✅A1--case_group is documented as required in arguments table
✅A2Validation fires at argument parsing before dependency check
✅A3Script exits with code 1
✅A4Error message identifies missing required argument
✅A515 SKILL_* codes cover missing/invalid parameter scenarios comprehensively
Pass rate: 5 / 5
Medical Task Total78.6 / 100
Key Strengths
- 15 SKILL_* error codes — the most comprehensive error coverage in the batch, including SKILL_PARSE_ERROR, SKILL_FILE_WRITE_ERROR, and SKILL_MEMORY_ERROR
- Strict less-than-150-lines-per-file modularity constraint with explicit statement in implementation checklist
- When to Read External Files table at the very top of SKILL.md enables rapid progressive disclosure for agents
- Natural trigger language with NOT for in frontmatter description prevents misuse for multiclass, survival, and linear regression