What problem does it solve? Teams using AI to generate Salesforce artifacts lose track of which prompts produced which components, making it impossible to trace generated code back to its origin or keep new artifacts consistent with established project conventions. ## Core Features & Use Cases - Convention inference before generation: Reads the existing .generated-prompts/<type>.md register before creating any artifact to infer naming conventions, data layer strategy, patterns, and sharing model decisions. - Prompt persistence after generation: Appends or updates one entry per component with creation date, iteration count, key decisions, and the final prompt, keeping only the latest version while git history holds the trail. - Typed register files: Covers Apex, LWC, flows, triggers, Visualforce, Aura, OmniStudio, metadata, diagrams, CI/CD, docs, and tests, each mapped to active team profiles. - Use Case: A developer asks the AI to create a new Apex service class; the Skill first reads .generated-prompts/apex.md to match the existing selector-layer pattern, then registers the new prompt entry after generation. ## Quick Start Before generating any new Salesforce component, read the matching file in .generated-prompts/ to follow existing conventions, then record the final prompt there after the artifact is created.
