contention-layer

What problem does it solve?

When two skills, agents, or research tracks produce conflicting outputs, valuable signal is often discarded as an error instead of being used to improve the ecosystem. This Skill eliminates that waste by analyzing contention to uncover new validation angles and generate new skill candidates for the Forge Harness ecosystem.

Core Features & Use Cases

• Conflict Classification: Categorizes conflicts into criteria, scope, order, philosophy, or dual-track grounding types to identify the root of disagreement between sources.
• Harvest Gate Validation: Runs a structured 3-question gate to determine if a conflict reveals a new, standalone skill opportunity or can be resolved by improving existing skills.
• Ecosystem Routing: Automatically routes new skill candidates to the correct Forge Harness path (fh-meta, commons plugin, or field harvest) based on their scope and domain.
• Track Conflict Handling: Specializes in dual-track grounding conflicts between open-frontier research and internal memory recall, resolving them via memory hygiene, publishable deltas, or phantom quench as appropriate. Use Case: If your deep-research track and internal CATALOG recall disagree on a factual claim, this skill verifies sources, identifies whether the conflict is a stale memory or a new publishable insight, and generates a new skill candidate if the disagreement reveals a novel validation angle.

Quick Start

Invoke the contention-layer skill when you notice two skills or research tracks producing conflicting conclusions about the same output to analyze the disagreement and route any new skill candidates to the appropriate Forge Harness ecosystem path.