Self Checks
1. Is this request related to a challenge you're experiencing? Tell me about your story.
Current Pain Point:
Currently, managing multiple skills (e.g., querying different internal APIs like ticketing systems, BI tools) requires manually building a complex router using prompt engineering and state machines (Conversation Variables).
As the number of skills grows (beyond 5-10), the dispatcher logic becomes extremely fragile and hard to maintain.
Inspired by "Lobster" Pattern:
We need a "Lobster" style capability where the AI can learn and automatically select the correct skill from a registry without manual routing configuration.
Proposed Feature:
Request for a native Skill Registry & Auto-Dispatcher mechanism.
- Skill Registry: A centralized place to define skills with structured JSON schemas (inputs/outputs).
- Auto-Matching: The AI runtime should automatically classify user queries and match them to the correct skill in the registry.
- Auto Slot-Filling: If required parameters are missing, the system should automatically manage the state and generate natural language follow-ups to collect information, instead of just stopping or giving errors.
- Seamless Execution: Once parameters are collected, execute the skill and return the result as part of the main conversation flow.
2. Additional context or comments
Use Case Example:
A user wants to analyze a network fault.
- User Input: "Check the status of ticket TICKET-1024"
- System Action: Matches "Check Ticket" skill -> Discovers missing "Department ID" -> Automatically asks "Which department does this belong to?" -> Runs analysis -> Returns result.
Reference:
This is often referred to in the community as the "Lobster" architecture (Single entry point, multiple backend skills, automatic chaining).
3. Can you help us with this feature?
Self Checks
1. Is this request related to a challenge you're experiencing? Tell me about your story.
Current Pain Point:
Currently, managing multiple skills (e.g., querying different internal APIs like ticketing systems, BI tools) requires manually building a complex router using prompt engineering and state machines (Conversation Variables).
As the number of skills grows (beyond 5-10), the dispatcher logic becomes extremely fragile and hard to maintain.
Inspired by "Lobster" Pattern:
We need a "Lobster" style capability where the AI can learn and automatically select the correct skill from a registry without manual routing configuration.
Proposed Feature:
Request for a native Skill Registry & Auto-Dispatcher mechanism.
2. Additional context or comments
Use Case Example:
A user wants to analyze a network fault.
Reference:
This is often referred to in the community as the "Lobster" architecture (Single entry point, multiple backend skills, automatic chaining).
3. Can you help us with this feature?