| name | question-parsing |
| description | Load and parse questions from /root/question.txt, extract question IDs and their artifact/product context. Validate that all questions are correctly mapped before proceeding to data retrieval. |
Question Parsing Skill
Step 1: Load Question File
- Read
/root/question.txt completely
- Handle any encoding issues (UTF-8)
- Preserve all question entries
Step 2: Parse Each Question
For each line/entry in the file:
- Extract the question ID (e.g., "q1", "q2", "q3")
- Extract the question text
- Identify the artifact/product context mentioned in the question
- Identify the entity type being asked for (e.g., reviewers, contributors, participants)
- Store parsed metadata in a structured dictionary
Step 3: Validate and Debug
- Print all parsed questions with their IDs, artifact contexts, and entity types
- Verify the total count matches expectations
- Check for any malformed entries
- Confirm key names and values are correctly extracted
Step 4: Return Parsed Questions
Return a dictionary mapping question IDs to:
{
"question_id": "q1",
"text": "full question text",
"artifact_context": "artifact or product name",
"entity_type": "reviewers|contributors|participants|etc",
"parsed_at": timestamp
}