automated-orchestration

What problem does it solve? Long-running data pipelines that fail mid-execution lose hours of progress, forcing full restarts. This Skill provides a checkpoint/resume architecture so every stage is resumable, idempotent, and adapts its analysis depth to the actual data volume. ## Core Features & Use Cases - Checkpoint/Resume Pipeline Stages: Each stage writes atomic JSON checkpoints, skips completed work on re-run, and resumes from partial progress after interruption or failure. - Activity-Based Depth Adjustment: Automatically selects shallow, standard, deep, or archival analysis tiers based on item count instead of hardcoding depth. - Rate-Limit-Aware Fetching: Exponential backoff with jitter for external API calls, plus partial-failure tolerance that logs errors without aborting the pipeline. - Use Case: Processing a large Reddit data export through fetch, parse, enrich, and report stages — if the enrich stage fails after 3 hours, re-running resumes from that stage's checkpoint instead of starting over. ## Quick Start Use the automated-orchestration skill to build a checkpointed pipeline that processes my data archive in resumable stages and writes a status report.