What problem does it solve? Coordinating multiple AI agents on a Salesforce feature often leads to file collisions, half-finished edits, unvalidated deploys, and broken work reaching production. This Skill structures Salesforce delivery as a five-wave workflow where parallel agents build on disjoint paths, machine-checked gates block failing work, and production deploys require validation before quick deploy. ## Core Features & Use Cases - Five-wave delivery model: scope and contract (wave 0), parallel build across Apex, LWC, metadata, and integration agents (wave 1), parallel checks (wave 2), serial validate-then-quick-deploy (wave 3), and post-deploy org verification (wave 4). - Collision-free parallelism: path ownership rules enforced by hooks so no two agents write the same file, with contracts fixing shared signatures before any code is written. - Blocking quality gates: vf-check commands for static analysis, Jest, Apex coverage, deploy validation, and smoke verification that must exit 0 before the next wave starts. - Use Case: A story asks to add renewal reminders to Opportunity, touching Apex, an LWC, a permission set, and a flow. The orchestrator scouts the impact, writes a contract, dispatches four engineers in one batch, gates the result, validates and quick-deploys to UAT, then verifies the feature in the org. ## Quick Start Ask the assistant to run the story workflow for your change, for example: run the five-wave Salesforce delivery workflow for the story "Add renewal reminders to Opportunity" starting with an impact map and contract.
