skill-phase-context

Optimize AI agent token usage by implementing a tiered loading strategy that categorizes skills into always loaded, phase-triggered, and on-demand tiers. This ensures only necessary skills are active, preventing infrequently used skills from consuming resources.

Tiered skill loading is a resource allocation strategy that organizes skills into TIER 0, TIER 1, and TIER 2. TIER 0 loads at session bootstrap, TIER 1 triggers on phase entry, and TIER 2 loads on explicit request.

Configure TIER 1 phase-triggered loading to load development-specific skills upon phase entry. This ensures that when an agent enters the Development phase, only relevant skills like developer-guidelines are activated rather than all available skills.

Yes, you can manage skill loading by assigning frequently needed skills to TIER 0 for session bootstrap and assigning infrequently used skills to TIER 2 for explicit on-demand loading. This optimizes workflow efficiency and resource allocation.

Organize skills using a tiered loading strategy to optimize performance. Categorize skills into bootstrap, phase-triggered, and on-demand tiers to ensure efficient resource allocation across multi-agent systems.

Loading all skills at once consumes excessive tokens because there is no tiered loading strategy. By defining TIER 0, 1, and 2 skill loading tiers, the agent only activates necessary skills at specific phases, reducing computational load.