Datasets:
id string | scenario string | dep float64 | buf float64 | lag float64 | cpl float64 | notes string | label_cascade_state int64 |
|---|---|---|---|---|---|---|---|
AI5DO-0001 | Critical dependencies have failover. Agent degrades to safe mode. On-call responds fast. | 0.42 | 0.82 | 0.12 | 0.4 | buffer strong | 0 |
AI5DO-0002 | Intermittent dependency errors. Retries enabled. Triage occurs daily. | 0.56 | 0.62 | 0.54 | 0.52 | lag rising | 0 |
AI5DO-0003 | Dependency latency spikes. Retry policy amplifies load. Alerts fire but response delayed. | 0.66 | 0.48 | 0.72 | 0.66 | outage risk | 1 |
AI5DO-0004 | Multiple tools depend on the same service. Failure propagates across workflows. | 0.74 | 0.4 | 0.78 | 0.78 | coupling tight | 1 |
AI5DO-0005 | Dependency outage triggers cascading tool failures. No pause checkpoint. Rollback delayed. | 0.82 | 0.34 | 0.82 | 0.82 | late intervention | 1 |
AI5DO-0006 | Circuit breakers and backoff engage automatically. Agent enters safe mode. Recovery fast. | 0.58 | 0.72 | 0.16 | 0.54 | fast response | 0 |
AI5DO-0007 | Throughput pressure disables backoff to reduce latency. Weekly review misses rising fragility. | 0.72 | 0.42 | 0.76 | 0.74 | buffer eroded | 1 |
AI5DO-0008 | New dependency added in sandbox. Chaos test run. Live monitor can pause rollout. | 0.46 | 0.78 | 0.14 | 0.46 | recoverable | 0 |
AI5DO-0009 | Shared dependency outage plus retry storms. Alert flood prevents triage for days. | 0.88 | 0.26 | 0.86 | 0.9 | cascade engaged | 1 |
What this repo does
This dataset models dependency outage cascades in tool-using AI systems. It detects when dependency stress rises, protective buffers weaken due to missing failover and backoff, governance lag delays triage, and tight coupling through shared dependencies propagates failures across workflows, crossing the five-node cascade threshold into an unrecoverable dependency outage cascade.
This dataset models a five-node cascade: four interacting instability drivers and one emergent cascade state.
The fifth node represents the nonlinear transition from recoverable drift to systemic cascade.
Core quad
dep
buf
lag
cpl
Prediction target
label_cascade_state
Row structure
One row represents an AI tool ecosystem scenario with numeric signals for dependency stress, safety buffer strength, governance lag, and coupling tightness through shared dependencies, paired with a cascade state label.
Files
data/train.csv
data/tester.csv
scorer.py
Evaluation
Run predictions on data/tester.csv and score with scorer.py.
License
MIT
Structural Note
This dataset identifies a measurable coupling pattern associated with systemic instability.
The sample demonstrates the geometry.
Production-scale data determines operational exposure.
What Production Deployment Enables
• 50K–1M row datasets calibrated to real operational patterns
• Pair, triadic, and quad coupling analysis
• Real-time coherence monitoring
• Early warning before cascade events
• Collapse surface and recovery window modeling
• Integration and implementation support
Small samples reveal structure.
Scale reveals consequence.
Enterprise & Research Collaboration
Clarus develops production-scale coherence monitoring infrastructure for critical systems across healthcare, finance, infrastructure, and regulatory domains.
For dataset expansion, custom coherence scorers, or deployment architecture:
team@clarusinvariant.com
Instability is detectable.
Governance determines whether it propagates.
- Downloads last month
- 18