What problem does it solve? In multi-agent workflows, agents publish vague or dishonest intent declarations that prevent overlap detection, conflict resolution, and convergence. This Skill enforces a structured schema and quality criteria so every intent published to Convergent's intent graph is specific, verifiable, and honest about its stability. ## Core Features & Use Cases - Intent Authoring: Creates well-structured IntentNode objects with specific actions, concrete provides/requires lists, declared constraints, and evidence-backed stability scores. - Intent Validation: Checks existing intents against an 8-point quality checklist covering specificity, evidence consistency, and circular dependency detection. - Intent Updates: Modifies intents when stability changes by ±0.2 or scope is refined, tracking dependents affected by the change. - Use Case: An agent choosing a data model publishes an interface intent declaring the User model fields, uniqueness constraints, and a stability score of 0.7 backed by passing tests, so other agents can safely build against it. ## Quick Start Ask the agent to author an intent for a decision you just made, providing the action, category, provides, requires, constraints, stability score, and affected files.
