End-to-end automation from CDP data extraction to Lucidchart prompt generation.
The AI agent retrieves Layer-2 topology data from AWX and parses CDP neighbor relationships. It passes the structured topology to Gemini AI to produce a Lucidchart-ready prompt. The result is a ready-to-paste diagram prompt that accurately reflects the network topology.
Core actions that transform data into a diagram-ready prompt.
Fetch CDP topology data from AWX and produce stdout.
Parse neighbor relationships using JavaScript to extract device links.
Transform parsed topology into a Lucidchart-ready prompt format.
Validate prompt structure for syntactic correctness.
Submit structured data to Gemini AI for final prompt generation.
Return a ready-to-paste Lucidchart prompt.
This AI agent replaces fragmented manual work with a predictable execution flow.
A simple 3-step flow any user can follow.
Trigger the AWX Job Template to run network commands (e.g., show cdp neighbors detail) and produce stdout.
n8n fetches stdout and uses a JavaScript parser to extract neighbor relationships and device links.
Send the structured data to Gemini AI to generate a Lucidchart-ready prompt.
One realistic scenario.
In a mid-size campus network of 12 devices, the AI agent triggers AWX, collects CDP neighbors, parses relationships, and returns a Lucidchart-ready prompt. This prompt yields an accurate diagram of inter-device connections that matches the physical layout. The process completes in under 6 minutes with logs available for auditing.
Roles that gain from automated topology prompts.
Requires quick, accurate topology visuals to implement changes.
Needs consistent, shareable topology prompts for designs.
Maintains up-to-date diagrams for run books and incident response.
Documents rack-level topology for cabling and replacements.
Represents hybrid networks in diagrams generated from real topology.
Visualizes L2 topology to assess attack paths and access controls.
The AI agent connects tools to automate the flow.
Runs the Job Template to gather CDP data and produces stdout read by the AI agent.
Fetches stdout, coordinates the parser, and passes structured topology to Gemini AI.
Generates a Lucidchart-ready prompt from the structured topology data.
Consumes the prompt to create the network diagram visually.
Concrete scenarios where this AI agent shines.
Common questions about using this AI agent.
The AI agent uses AWX to execute CDP-relevant commands and capture stdout, then parses the output to extract neighbor relationships. It relies on a defined topology schema to ensure consistency. All parsing is performed in a controlled pipeline with auditable steps. If sources change or commands fail, the agent surfaces a clear error with suggested remediation.
Data is processed within your environment through AWX and your chosen LLM provider. Access to AWX and the AI tool is governed by your credentials and role-based access controls. The agent does not transmit sensitive data outside your network unless you configure it to do so. Output prompts are stored temporarily for the duration of the flow and can be purged on request.
The end-to-end flow typically completes in a few minutes, depending on network size and AWX job duration. The AWX step runs in parallel with the parsing stage, and Gemini AI completes the final prompt generation quickly once structured data is available. Logs provide visibility into timing and any bottlenecks for optimization. You can re-run the flow to regenerate prompts after topology changes.
Yes. The AI agent uses a defined prompt structure that you can modify to include or exclude specific fields. You can adjust how devices are labeled, what relationships are emphasized, and how the diagram should be organized (e.g., by site or device type). Customization can be saved as a template and reused for future topologies. Any changes are reflected in subsequent prompt generations.
If the AWX job fails, the AI agent halts the flow and surfaces a detailed failure report. It includes the command that failed, the stdout excerpt, and suggested remediation steps (e.g., retry, credentials check, or Job Template ID validation). You can re-run after addressing the root cause. The prompt generation step only executes after a successful AWX run to prevent cascading errors.
The agent is designed to operate on the topology data returned by CDP, which can include multiple sites. It can handle heterogeneous environments if the CDP data is available. For multi-vendor environments, ensure CDP equivalents or vendor-specific neighbor discovery data are used. The output prompt will reflect the topology as parsed and organized by the defined schema.
The current prompts are built from CDP-derived data, which commonly focuses on IPv4, but the parser can be extended to include IPv6 neighbor information if available. Gemini AI can incorporate IPv6 labeling in the Lucidchart prompt. If IPv6 data is critical, you can adjust the data collection commands and parsing rules accordingly. Expect prompts to represent IPv6 addresses in the same layout as IPv4 where provided.
End-to-end automation from CDP data extraction to Lucidchart prompt generation.
