context-minimizer

To generate an inline dispatch bundle, assemble a self-contained Markdown payload containing role contracts, governance excerpts, and artifact inputs directly into the task prompt. This eliminates pre-dispatch file reads by embedding all necessary context inline.

Passing project context safely requires assembling a per-dispatch bundle that respects token ceilings and embeds directly into the task prompt. It uses the project context cache and falls back to live extraction from PROJECT_CONFIG.md as needed.

Project context cache works during task dispatch by providing pre-assembled domain context for per-role bundle assembly. When the cache is unavailable, the system gracefully falls back to live extraction from PROJECT_CONFIG.md to ensure the target agent receives context.

You need a PROJECT_CONFIG.md file to assemble governance and domain context when the project context cache is unavailable. The system uses this live extraction fallback mechanism to read role contracts and ensure the inline dispatch bundle contains required artifacts.

Yes, you can include role contracts and governance excerpts in a single inline Markdown payload. The dispatch bundle assembles these components alongside project context sections and artifact inputs, creating a self-contained payload embedded directly in the task prompt.

Inline dispatch bundles replace pre-dispatch file reads by embedding governance, domain context, and artifact inputs directly into the task prompt. This ensures the target agent receives a self-contained payload that fits within token ceilings during AI workflow orchestration.