a5c-ai/babysitter / library/specializations/data-science-ml/skills/alibi-explainer
alibi-explainer
VerifiedAlibi explainability skill for counterfactual explanations, anchors, and trust scores.
README.md
View on GitHub# alibi-explainer
Alibi explainability skill for counterfactual explanations, anchors, and trust scores.
## Purpose
This skill provides advanced model explainability using Alibi, enabling teams to generate counterfactual explanations, anchor rules, and trust scores for complex models.
## Key Features
- **Counterfactuals**: "What-if" explanations showing minimal changes for different outcomes
- **Anchors**: High-precision rule-based explanations
- **Trust Scores**: Confidence measures for predictions
- **CEM**: Contrastive explanations with pertinent positives/negatives
- **Integrated Gradients**: Attribution for neural networks
## When to Use
- Explaining why a specific outcome occurred and what could change it
- Generating rule-based explanations for stakeholders
- Measuring prediction confidence
- Regulatory compliance requiring contrastive explanations
- Debugging edge cases and model failures
## Integration
Works seamlessly with:
- `shap-explainer` for feature attribution
- `lime-explainer` for local explanations
- `fairlearn-bias-detector` for fairness analysis
- `model-card-generator` for documentation
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |