growthxaiOfficial·33 Agent Skills Included

output

Build, debug, and evaluate durable LLM workflows

Builds production-ready LLM workflows with prompts, evals, tracing, cost tracking, and encrypted credentials in one TypeScript framework. Eliminates scattered SaaS subscriptions, fragile prompt dashboards, and manual retry logic by baking in durable Temporal orchestration. Guides coding agents through planning, scaffolding, debugging, and evaluating workflows with ready-made skills and error-fixing playbooks.

430|12

↗

npx skills add growthxai/output --all -g -y

Available:

Tells the AI agent how the Output framework is structured, which specialized sub-agents to call for each technical domain, and which coding and validation rules to follow when building workflows.

All Skills in This Repository (33)

Pure Emerald Level Indicators

📦 In Repo

growthxai

output-workflow-stop

Stop running Output SDK workflows by ID via the npx command.

output-error-direct-io

Move direct I/O operations in Output SDK workflows into step functions.

output-error-nondeterminism

Identify and fix non-deterministic patterns in Output SDK workflows.

output-credentials-env-vars

Resolve credential: references into environment variables at worker startup.

output-error-zod-import

Replace direct zod imports with @outputai/core in Output SDK workflows.

output-workflow-trace

Analyze Output SDK workflow traces to identify failures and performance bottlenecks.

output-meta-post-flight

Audit Output SDK workflows for step execution, conventions, and deliverables.

output-workflow-result

Retrieve completed Output SDK workflow results for post-run analysis.

output-workflow-run

Execute Output SDK workflows synchronously with inline or file-based JSON inputs.

output-dev-create-skeleton

Generates a standardized skeleton for new Output workflow projects.

output-error-missing-schemas

Detect missing inputSchema/outputSchema definitions in Output SDK workflows and guide fixes with examples.

output-error-try-catch

Remove try-catch blocks around Output SDK step calls to enforce retry policies.

Official

Intermediate

Frequently Asked Questions

FAQPage Schema

How to install Output?▼

Run `npx skills add growthxai/output --all -g -y` in your terminal to install all skills in this suite globally.

What does the Output framework do?▼

It is an open-source TypeScript framework for building reliable LLM workflows and agents with prompts, evals, tracing, cost tracking, and encrypted credentials built in. Temporal runs underneath, so failed workflows retry and resume automatically.

How do I debug a failing Output workflow?▼

The included skills walk your agent through checking services, listing runs, reading execution traces, and matching errors to known fixes like wrong Zod imports or direct I/O in workflow code.

Does Output work with Claude Code?▼

Yes. The framework is designed for AI coding agents, and every skill follows the standard SKILL.md format that works in Claude Code, Cursor, and similar tools.

Can I test LLM output quality automatically?▼

Yes. Output includes an eval system with LLM-as-judge evaluators, dataset-driven offline tests, and skills for error analysis, judge validation, and eval suite audits.