trigger-cost-savings

trigger-cost-savings analyzes Trigger.dev tasks, schedules, and runs to find cost-optimization opportunities. It solves the problem of runaway or over-provisioned background-job spend by combining static source-code analysis with live run analysis, then producing a prioritized report of concrete, estimated-impact recommendations. It is the skill to reach for when asked to reduce spend, audit usage, right-size machines, or review task efficiency.

The workflow has three parts. Static analysis scans task files for issues such as oversized machine presets, missing maxDuration, excessive retries, missing debounce or idempotency, polling loops instead of checkpointed waits, short waits, sequential triggers that could be batched, and over-scheduled crons. Run analysis uses the Trigger.dev MCP server tools (list_runs, get_run_details, get_current_worker) to identify expensive tasks, analyze failure patterns and wasted retries, check machine utilization versus actual resource use, and review schedule frequency; if the MCP tools are not available the skill instructs the user to install them via `npx trigger.dev@latest install-mcp` and can still run static analysis. It also fetches the latest official cost-reduction documentation via WebFetch, falling back to a bundled references/cost-reduction.md. Recommendations are presented as a prioritized report with high/medium/low impact and code snippets.

The target users are engineering teams and platform owners running workloads on Trigger.dev who want to lower compute costs without changing business behavior. All analysis is read-only: it inspects source and run data and fetches public docs, with no destructive or data-exfiltration behavior.