Agent skill · marketing growth · jeremylongshore
lindy-performance-tuning
Optimize Lindy AI agent execution speed, reliability, and cost efficiency.
Why this skill is useful
Provides specific optimization techniques and strategies for improving Lindy AI agent performance that the AI wouldn't reliably generate on its own.
What it needs
Requires lindy account access. About 5k tokens when loaded. Last updated 2026-08-07. 2,604 stars on the source repository.
What this skill does
Lindy Performance Tuning Overview Improve a Lindy workflow through controlled, workspace-specific experiments. Measure the current workflow in Tasks, preserve approval and security boundaries, change one variable, run the same sanitized fixture/eval cohort, and keep the change only when predeclared latency, quality, reliability, and cost criteria pass. Prerequisites Lindy workspace with active agents Access to agent Tasks tab (view step-by-step execution history) A representative, data-minimized fixture set and offline eval cohort Permission to inspect Tasks and restore a saved version An owner-approved list of invariants: confirmations, authorization, redaction, retention, and side-effect limits Instructions Step 1: Define the Experiment Contract Before editing, record: one workflow class and representative sanitized fixtures; the exact variable to change and why it should help; baseline sample window and sample count; duration median/p95, outcome ratio, eval score, and workspace-reported usage/cost; acceptance and rollback thresholds; and security/approval invariants that must remain unchanged. Use the Tasks view and Get Task Details to identify the slowest or least reliable block. Do not infer universal latency or credit values from this skill. Step 2: Save a Rollback Point Open Version History and identify the last known-good version. Record its timestamp or label. Restoring a version creates a new editable version; it does not erase the current history. Keep production activation and side-effecting actions unchanged until the candidate passes the evaluation lane. …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills jeremylongshore/lindy-performance-tuning