get_glossary_conflict_guide

Returns a step-by-step recipe for detecting conflicts between the AI agent’s glossary and the rest of its configuration — custom instructions, company description, coaching, and knowledge articles. A conflict is anything elsewhere in the configuration that contradicts, duplicates, or undermines a glossary term, so the agent would behave inconsistently.

The recipe is read-only: it tells the assistant which tools to call and how to classify and report what it finds, but it never changes any configuration.

Call this when asked to audit, review, or check the glossary for conflicts, duplicates, or contradictions — before reading the glossary, so the assistant follows the recommended workflow rather than improvising.

Example prompts

• “Check my glossary for conflicts with my custom instructions.”
• “Audit the glossary — are any terms contradicted elsewhere in my configuration?”
• “Is the term group-plan consistent with the rest of my Agent’s setup?”

Parameters

Response

Returns markdown describing how to:

• Load the glossary with list_glossary_terms (paginating until every term is fetched).
• Load the other configuration with get_ada_configuration (company description, custom instructions, playbooks, web actions).
• Cross-reference each term against coaching and knowledge using search_coaching and search_knowledge.
• Classify conflicts into four types (behaviour defined elsewhere, heavy term usage, intra-glossary overlap, contradictory definitions).
• Report findings in a structured Glossary Conflict Report with a suggested resolution per conflict.