Agent skill · data analytics · jeremylongshore
lindy-observability
Monitor Lindy AI agent health, task success rates, and credit consumption.
Why this skill is useful
Adds specific webhook configurations and metrics collection patterns for monitoring Lindy AI agent performance that the AI wouldn't reliably generate on its own.
What it needs
Requires express, prom-client installed locally. About 5k tokens when loaded. Last updated 2026-08-07. 2,604 stars on the source repository.
What this skill does
Lindy Observability Overview Monitor workflow health from Lindy's documented task surfaces. Start with Tasks for manual inspection, then use an Agent Task Change trigger followed by Get Task Details for workflow-based monitoring and send only bounded operational fields to an external collector; task inputs, outputs, customer content, and secrets do not belong in metrics or logs. Prerequisites Lindy workspace with active custom agents Access to each monitored agent's Tasks view For external monitoring: an HTTPS receiver and a metrics stack A distinct, nonempty callback secret stored as LINDYCALLBACKSECRET by the receiver and as a protected value in the Lindy HTTP Request action Authentication and Data Boundary Authenticate Lindy's outbound HTTP Request with a dedicated bearer value generated for the metrics receiver. Store it only in Lindy's protected action configuration and the receiver's secret manager, require at least 32 characters, compare it in constant time, and rotate it independently. Never reuse an inbound Lindy webhook secret or a metrics-scrape credential. Export only the three schema fields defined below. Instructions Step 1: Establish the Built-In View 1. Open the custom agent and select Tasks. 2. Review task status and open representative runs. 3. Inspect chronological steps, timestamps, conditions, and the error location. 4. Record a workspace-specific baseline by agent and workflow class. Do not copy task inputs or outputs into the baseline. The documented sources for operational signals are: Signal Source Handling --- --- --- Task outcome and frequency Tasks / Agent Task Change Aggregate by configured agent key Duration and failing block Get Task Details Retain duration; keep block content in Lindy Workspace spend Lindy billing view Keep billing data at its documented source Step 2: Build the Monitoring Workflow Create a separate monitoring agent using documented Lindy utilities: 1. Add Agent Task Change as the trigger. 2. …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills jeremylongshore/lindy-observability