benjamin-blanchet

Run focused corpus passes. The Corpus agent executes all passes; maturity changes the mode, not the agent.

Audit the corpus for deployability and maintainability.

Cross-check corpus against code, prod, tickets and knowledge-base sources.

Maintain the persistent roadmap and repo-native knowledge graph used by continuous corpus enrichment.

Check whether a continuous corpus run actually increased durable knowledge, updated state honestly and left a useful next step.

Keep the operator in control during corpus initialization and continuous enrichment by ending important `Corpus` responses with a compact…

Capture durable knowledge into the corpus after a task, discussion, incident, spec or implementation.

Run deterministic checks on the application corpus and turn the results into concrete maintenance actions.

Aggregate everything that crosses feature boundaries: the API surface, the domain model, the integration map, the **messaging…

Make corpus kickstart behave like a multi-disciplinary analysis team, not a light documentation generator.

Make corpus kickstart coverage explicit, measurable and hard to fake.

Provide reusable recipes for common continuous corpus runs.