Data Analysis
scvi-tools
86100Total Score
Core Capability
84 / 100
Functional Suitability
11 / 12
Reliability
9 / 12
Performance & Context
7 / 8
Agent Usability
14 / 16
Human Usability
8 / 8
Security
10 / 12
Maintainability
9 / 12
Agent-Specific
16 / 20
Medical Task
20 / 20 Passed
91Deep generative models for single-cell omics; use when you need probabilistic batch correction (scVI), transfer learning, uncertainty-aware differential expression, or multimodal integration (totalVI/MultiVI)
4/4
87Deep generative models for single-cell omics; use when you need probabilistic batch correction (scVI), transfer learning, uncertainty-aware differential expression, or multimodal integration (totalVI/MultiVI)
4/4
85Unified model API: setup_anndata(...) → Model(adata) → train() → get_*() across model families
4/4
85Probabilistic latent representations for integration, denoising, and downstream clustering/visualization
4/4
85End-to-end case for Unified model API: setup_anndata(...) → Model(adata) → train() → get_*() across model families
4/4
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 scientific-integrity problem was surfaced because the package did not claim more than the available records, article text, or script evidence supported. |
| Practice Boundaries | PASS | The evaluated outputs stayed inside the Deep generative models for single-cell omics and did not drift into unsupported interpretation beyond the available inputs. |
| Methodological Ground | PASS | The legacy review kept the package aligned with its named analysis library, data structure, or processing workflow. |
| Code Usability | PASS | The archived review preserved a usable code path with named scripts, expected inputs, and a recognizable output contract. |
Core Capability84 / 100 — 8 Categories
Functional Suitability
Functional suitability was softened by the legacy issue 'Improve stress-case output rigor'. Stress and boundary scenarios show weaker consistency
11 / 12
92%
Reliability
Related legacy finding for scvi-tools: Improve stress-case output rigor. Stress and boundary scenarios show weaker consistency
9 / 12
75%
Performance & Context
The package performed well overall, with only a small remaining performance-context deduction.
7 / 8
88%
Agent Usability
The archived review left some headroom in how quickly an agent can lock onto the intended analysis path.
14 / 16
88%
Human Usability
No point loss was recorded for human usability in the legacy audit.
8 / 8
100%
Security
The packaged workflow stayed safe overall, with only a small remaining deduction around boundary signaling.
10 / 12
83%
Maintainability
The analysis package is maintainable overall, though the archived score suggests modest cleanup headroom.
9 / 12
75%
Agent-Specific
Related legacy finding for scvi-tools: Improve stress-case output rigor. Stress and boundary scenarios show weaker consistency
16 / 20
80%
Core Capability Total84 / 100
Medical TaskExecution Average: 86.6 / 100 — Assertions: 20/20 Passed
91
Canonical
Deep generative models for single-cell omics; use when you need probabilistic batch correction (scVI), transfer learning, uncertainty-aware differential expression, or multimodal integration (totalVI/MultiVI)
4/4 ✓
87
Variant A
Deep generative models for single-cell omics; use when you need probabilistic batch correction (scVI), transfer learning, uncertainty-aware differential expression, or multimodal integration (totalVI/MultiVI)
4/4 ✓
85
Edge
Unified model API: setup_anndata(...) → Model(adata) → train() → get_*() across model families
4/4 ✓
85
Variant B
Probabilistic latent representations for integration, denoising, and downstream clustering/visualization
4/4 ✓
85
Stress
End-to-end case for Unified model API: setup_anndata(...) → Model(adata) → train() → get_*() across model families
4/4 ✓
91
Canonical✅ Pass
Deep generative models for single-cell omics; use when you need probabilistic batch correction (scVI), transfer learning, uncertainty-aware differential expression, or multimodal integration (totalVI/MultiVI)
Deep generative models for single-cell omics; use when you need... remained tied to the documented analysis contract even when the preserved evidence centered on instructions instead of a full rerun.
Basic 36/40|Specialized 55/60|Total 91/100
✅A1The scvi-tools output structure matches the documented deliverable
✅A2The instruction path remains actionable for the documented case
✅A3The output stays fully within the documented skill boundary
✅A4The response quality is acceptable for the documented path
Pass rate: 4 / 4
87
Variant A✅ Pass
Deep generative models for single-cell omics; use when you need probabilistic batch correction (scVI), transfer learning, uncertainty-aware differential expression, or multimodal integration (totalVI/MultiVI)
Deep generative models for single-cell omics; use when you need... remained tied to the documented analysis contract even when the preserved evidence centered on instructions instead of a full rerun.
Basic 34/40|Specialized 53/60|Total 87/100
✅A1The scvi-tools output structure matches the documented deliverable
✅A2The instruction path remains actionable for the documented case
✅A3The output stays fully within the documented skill boundary
✅A4The response quality is acceptable for the documented path
Pass rate: 4 / 4
85
Edge✅ Pass
Unified model API: setup_anndata(...) → Model(adata) → train() → get_*() across model families
Unified model API: setup_anndata(...) → Model(adata) → train() →... remained tied to the documented analysis contract even when the preserved evidence centered on instructions instead of a full rerun.
Basic 33/40|Specialized 52/60|Total 85/100
✅A1The scvi-tools output structure matches the documented deliverable
✅A2The instruction path remains actionable for the documented case
✅A3The output stays fully within the documented skill boundary
✅A4The response quality is acceptable for the documented path
Pass rate: 4 / 4
85
Variant B✅ Pass
Probabilistic latent representations for integration, denoising, and downstream clustering/visualization
The archived run treated Probabilistic latent representations for integration, denoising,... as a bounded analysis workflow rather than a purely narrative instruction path.
Basic 32/40|Specialized 53/60|Total 85/100
✅A1The scvi-tools output structure matches the documented deliverable
✅A2The instruction path remains actionable for the documented case
✅A3The output stays fully within the documented skill boundary
✅A4The response quality is acceptable for the documented path
Pass rate: 4 / 4
85
Stress✅ Pass
End-to-end case for Unified model API: setup_anndata(...) → Model(adata) → train() → get_*() across model families
The archived run treated End-to-end case for Unified model API: setup_anndata(...) →... as a bounded analysis workflow rather than a purely narrative instruction path.
Basic 29/40|Specialized 56/60|Total 85/100
✅A1The scvi-tools output structure matches the documented deliverable
✅A2The instruction path remains actionable for the documented case
✅A3The output stays fully within the documented skill boundary
✅A4The response quality is acceptable for the documented path
Pass rate: 4 / 4
Medical Task Total86.6 / 100
Key Strengths
- Primary routing is Data Analysis with execution mode A
- Static quality score is 84/100 and dynamic average is 78.6/100
- Assertions and command execution outcomes are recorded per input for human review
- Execution verification summary: No script verification was applicable