
from ai-plugin52
Detects silent delivery failures across data pipelines (CDP functions, batch exports, and flows) and surfaces high-confidence findings or durable memory entries
Signals Scout — Data Pipelines audits a project's data delivery surfaces (CDP destinations and transformations, batch exports, and hog flows) to find contradictions between configured state and actual delivery. It identifies watcher interventions (auto-disabled or degraded functions), failure-share spikes for destinations, stalled or failed batch export runs, and concentrated flow failures, then emits high-confidence signals or writes durable memory for lower-confidence observations.
Use this skill when you need automated monitoring and triage of downstream delivery problems that are often silent: missing exports, destinations that stopped delivering, or flows failing for recipients. It is intended for periodic scout runs (daily or on-demand) to surface P1/P2 delivery issues and to avoid noisy duplicate reports by deduping against prior findings.
Designed for PostHog Signals agents (Claude sandbox) with read-only analytics and signals-scout write scopes. Compatible with MCP tools like cdp-functions-list, batch-export-get, workflows-global-stats.
This skill has not been reviewed by our automated audit pipeline yet.