| name | config-creator |
| description | Skill Creator — Autonomously onboards external APIs by discovering schemas, generating platform-native tools, and deploying schema-aware workspaces |
| metadata | {"version":"3.0.0","author":"Network AI Team","type":"agent","category":"system-integration","intent":"skill_deployment_external_system_onboarding"} |
| tools | ["list_platform_services","read_api_schema","create_service_config","register_api_service","extract_schema_reference","create_skill_workspace","read_file","write_file"] |
| system | $ref:./prompts/system.md |
Overview
The Skill Creator autonomously onboards external APIs into OLAV.
It discovers API schemas, uses the platform's tool generation pipeline to produce
production-quality per-endpoint tools, and deploys a ready-to-use workspace —
all without hardcoding and without human intervention.
Workflow
- Explore —
list_platform_services() → understand current state
- Analyze —
read_api_schema(url) → discover endpoints, tags, auth
- Configure —
create_service_config(...) → write services.yaml entry
- Generate —
register_api_service(name) → platform builds per-endpoint tools
- Deploy —
create_skill_workspace(...) → workspace + SKILL.md + PLATFORM.md
- Verify —
read_file(generated_tool_path) → confirm quality
Design Principles
- LLM-Native: tag groups, auth type, workspace structure all inferred from schema semantics
- No hardcoding: tools are generic; LLM decides all service-specific parameters
- Platform pipeline: generated tools use
service_call() — never write raw Python code