dmzoneill InScope Research
Query Red Hat InScope AI assistants for internal documentation.
Inputs
| Input | Type | Default | Purpose |
|---|---|---|---|
question |
string | required | Question to ask |
assistants |
string | - | Comma-separated: app-interface, clowder, konflux. Empty = auto. |
refresh_auth |
bool | false | Force re-auth before query |
Persona
persona_load("developer")— InScope tools
Workflow
1. Load Persona
persona_load("developer")
2. Check Auth
inscope_auth_status()— check token validity- If expired or
refresh_auth:inscope_auto_login(headless=true)
3. List Assistants (optional)
inscope_list_assistants()— available assistants
4. Query
- If no
assistants:inscope_ask(query=inputs.question, include_sources=true)— auto-select - If
assistants:inscope_query(query=inputs.question, assistant=first_assistant, include_sources=true)
5. Parse Results
- Extract answer, sources, assistant used
- If unauthorized:
learn_tool_fix("inscope_ask", "unauthorized", "Token expired", "Run inscope_auto_login()") - If timeout:
learn_tool_fix("inscope_ask", "timeout", "Request timed out", "Retry with simpler question")
6. Log
memory_session_log("InScope research", "Question: {query[:80]}, Assistant: {assistant}")
MCP Tools
inscope_ask,inscope_query,inscope_list_assistantsinscope_auth_status,inscope_auto_login,inscope_save_token
Quick Examples
skill_run("inscope_research", '{"question": "How do I configure RDS for my ClowdApp?"}')
skill_run("inscope_research", '{"question": "Konflux release process", "assistants": "konflux"}')