home / skills / willsigmon / sigstack / glif-expert

glif-expert skill

safe

/plugins/design-tools/skills/glif-expert

This skill helps you design and run Glif AI workflows by guiding API integration, workflow graph creation, and app embedding.

npx playbooks add skill willsigmon/sigstack --skill glif-expert

Review the files below or copy the command above to add this skill to your agents.

Files (1)

SKILL.md

939 B

---
name: Glif Expert
description: Build and run Glif AI workflows - API integration, workflow design, app embedding
allowed-tools: Bash, Read, Edit, WebFetch
---

# Glif Expert

Complete Glif workflow expertise.

## API Usage
**Base URL**: `https://simple-api.glif.app`
**Auth**: Bearer token from https://glif.app/settings/api-tokens

```bash
curl -X POST https://simple-api.glif.app \
  -H "Authorization: Bearer $GLIF_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"id": "glif-id", "inputs": {"prompt": "..."}}'
```

## Workflow Design
Glifs are JSON graphs with blocks connected via variable references.

**Block types**: Text input, image input, GPT-4, Claude, DALL-E, Stable Diffusion, code execution

## Integration Patterns
- **Node.js**: `@glifxyz/sdk` package
- **Discord bots**: Webhook triggers
- **Web apps**: REST API + webhooks for async

Use when: Creating glifs, running AI workflows, integrating Glif into apps

Overview

This skill packages hands-on expertise for building and running Glif AI workflows, including API usage, workflow design, and app embedding. It covers authentication, request patterns, common block types, and integration tips for Node.js, web apps, and bots. The content is practical and focused on getting Glif workflows running reliably in production.

How this skill works

Glif workflows are modeled as JSON graphs where blocks (text, image, LLMs, diffusion models, code) connect via variable references. You trigger runs by POSTing workflow JSON to the Glif API with a Bearer token, and receive synchronous or webhook-driven asynchronous responses. SDKs and webhooks simplify integration into Node.js services, Discord bots, and web frontends for both immediate and long-running tasks.

When to use it

Orchestrating multi-step AI tasks that combine models and custom code
Embedding AI workflows into web apps or dashboards with REST + webhooks
Running automated content pipelines: generation, revision, and image creation
Connecting Glif flows to chatbots or community bots (e.g., Discord)
Prototyping production-ready AI features without building orchestration from scratch

Best practices

Store and rotate API tokens securely; use environment variables and secrets management
Design workflows as small, composable blocks to simplify testing and reuse
Use webhooks for long-running jobs and implement retry/idempotency on receivers
Validate inputs and sanitize model outputs before persisting or exposing them
Log block execution and errors with contextual IDs for fast debugging

Example use cases

Customer support automation: chain intent detection, response generation, and action execution
Creative pipelines: generate prompts, produce images with diffusion models, and post-process results
SaaS feature embedding: call Glif from a web app to provide on-demand writing, code, or image generation
Discord moderation bot: trigger Glif via webhooks to analyze messages and produce suggested actions

FAQ

How do I authenticate requests?

Use a Bearer token from your Glif account settings and include it in the Authorization header for API calls.

Which SDK should I use for Node.js?

Use the @glifxyz/sdk package for convenient request and webhook helpers in Node.js projects.

How do I handle long-running workflows?

Start runs via the REST API and provide a webhook endpoint to receive async completion events; implement retries and idempotency.