Draw data flows with openhop and AI
Describe a sequence in YAML and inspect it in a local step-through renderer, with messages, example data and expandable sub-flows.
Draw data flows with openhop and AI when the order of interactions matters as much as the components. This skill turns an authorised description into YAML for a companion local renderer, where readers can play the sequence and inspect what crosses each connection.
Show order submission, stock checking and payment response as separate steps, for example. A complex part can open into its own sub-flow. Keep the editable source, but expect to run the local service for viewing; this workflow does not promise a standalone HTML export or automatic publication.
What it does
- Explain a request across several services
- Show the message and data at each step
- Prepare a technical walkthrough for discussion
- Separate component categories from project names
- Open a complex section as a nested flow
- Revise a diagram through its YAML source
Install and use
- Provide the skill and its same-publisher npm runtime to a file-and-shell-capable agent. Node 22.12+ and a browser form one supported route. Local use needs no service account; setup also writes agent skill files.
- Describe components, sequence, messages and authorised example data. Distinguish a stock failure from successful payment rather than presenting one idealised path.
- Ask the agent to validate the YAML, submit it to the local service and return that flow’s viewing link. Check order, labels, data and nested navigation in the browser.
- Keep revisions in the source and the service running while needed. Do not expose private diagrams publicly; selected icons and fonts can make separate network requests.
FAQ
Source
Made by Naor Sabag. Original skill: openhop, published on GitHub under the MIT licence.
View on GitHub