Software-Defined Control Switch Diagnostic Agent with Claude Agents SDK and Metaflow
Critical industrial network switch failures corrupt control logic communication and impair operator situational awareness. Build an automated network diagnostic pipeline using Claude Agents SDK and Metaflow workflows to trace packet loss anomalies to exact control loop failures within 10 seconds.
What you are building
The core problem, expected build, and operating context for this challenge.
Develop an automated pipeline for diagnosing digital communications failures in software-defined controls using Claude Agents SDK and Metaflow.
How work is evaluated
Evaluates accuracy of identifying the root switch port failure from corrupted SDC telemetry.
Challenge datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.
port_match
Checks if failed port is correctly identified.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
loop_identification_recall
Recall metric of impacted control loops identified • target: 0.95 • range: 0-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Build a Metaflow graph to ingest high-frequency switch port telemetry
Utilize Claude Agents SDK with extended thinking to infer broken control loop dependencies
Correlate lost heartbeats with operator HMI alarm flooding
Generate automated fallback routing plans for Software-Defined Controls (SDC)
Reference links and supporting material
Network packet loss and port CRC error telemetry streams from redundant ring control networks.
Challenge setup
Evaluates accuracy of identifying the root switch port failure from corrupted SDC telemetry.
Challenge input
JSON object with network_topology and port_stats list
Claude Agents SDK
Anthropic agent framework for advanced extended thinking and reasoning
Metaflow
Framework for defining, executing, and monitoring automated data flows
Evaluated output
JSON object with failed_port (string), affected_control_loops (array), recovery_action (string)
- Checks if failed port is correctly identified.
- Recall metric of impacted control loops identified • target: 0.95 • range: 0-1
- Benchmark: SDC-NetDiag-2025
- Loop Identification Recall target: 0.95
- 1 public reference case
- Python execution harness
- Python sandbox (unavailable on Versalist)
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
- Metaflow · Required
- Claude Code · Optional
- Claude Code · Optional
- Claude 4.1 Opus · Optional
Evaluation contract
- port_match · Weight 1
- loop_identification_recall · Weight 1
Metrics
- loop_identification_recall · Target 0.95
Benchmarks
- SDC-NetDiag-2025
- security
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON object with network_topology and port_stats list
Output: JSON object with failed_port (string), affected_control_loops (array), recovery_action (string)
Execution harness: Python
Evaluation tools
- Claude Agents SDK · agent_framework
- Metaflow · ai_workflow_automation
Recipe freeze marker
No recipe freeze time is recorded.