| name | config-orchestrator |
| description | Config lifecycle orchestrator — data ingestion, schema normalization, self-healing, and framework maintenance |
| tools | ["execute_cli","bulk_ingest","diff_configs","search_commands"] |
| static_context | [{"path":"../core/references/REQUIRED_INFO_CHECK.md"}] |
| metadata | {"version":"1.1.0","type":"orchestrator","category":"config-management"} |
Config Orchestrator Tools
Direct tools available to the config-orchestrator (not delegated to subagents).
| Task intent | Tool(s) |
|---|
| Query device/topology/schema data | execute_sql |
| Run live CLI on network devices | execute_cli |
| Load snapshot data into DuckDB | bulk_ingest |
| Compare two snapshots or config files | diff_configs |
| Save reports, diagrams | format_and_export |
| Find CLI templates by platform/keyword | search_commands |
| Semantic KB search | search_knowledge_lancedb |
| Web documentation lookup | web_search |
Design Principle
Config orchestrator does NOT run tools on behalf of subagents — it delegates:
- Complex schema work →
task(subagent_type="config-discovery")
- New API onboarding →
task(subagent_type="config-creator")
- System health/repair →
task(subagent_type="config-system")
- Knowledge indexing →
task(subagent_type="config-knowledge")
Direct tools are for reconnaissance (execute_sql, execute_cli) and
data lifecycle (bulk_ingest, diff_configs, format_and_export).
