Workflows
Tech Story Teller consists of four modular n8n workflows. Each can be imported directly into n8n and chained together or triggered independently.Workflow A: GitHub Webhook Receiver
Purpose: Receives GitHub Action webhooks, extracts file paths, fetches content, and forwards to Workflow B.Flow
Nodes
| Node | Type | Function |
|---|---|---|
| Webhook Receiver | webhook | Receives POST with file list |
| Parse File Paths | function | Extracts file array from payload |
| Fetch Rust File | httpRequest | Calls GitHub API for file content |
| Decode Content | function | Base64 decodes file content |
| Send to Workflow B | executeWorkflow | Forwards to structural analysis |
Workflow B: Structural Analysis
Purpose: Takes a Rust file and sends it to LLM for AST-like extraction.LLM Task
The LLM analyzes the Rust source and outputs structured JSON:Configuration
- Model: gpt-4o-mini
- Temperature: 0 (deterministic output)
Workflow C: Deep LLM Analysis
Purpose: Generates comprehensive documentation from the structural analysis.Output Components
- ELI5 explanation of the file
- Senior-system-engineer explanation
- Function-by-function explanation
- Real-world analogies
- Rust concepts used
- Component interaction map
- Mermaid diagram showing file relationships
Configuration
- Model: gpt-4o
- Temperature: Default (balanced creativity)
Workflow D: Test Coverage + Report
Purpose: Verifies tests, identifies gaps, and assembles the final report.Tasks
- Run
cargo test -- --nocapture - Identify missing unit tests
- Map functions to existing tests
- Suggest fixes for failures
- Recommend high-value tests per function
- Produce final structured report
Output Delivery
- Slack: Webhook integration
- Notion: Page creation (configurable)
- Email: SMTP integration (configurable)
Chaining Workflows
For full automation, chain all four workflows:executeWorkflow nodes to pass data to the next stage.