meld
Version:
Meld: A template language for LLM prompts
46 lines (27 loc) • 904 B
Markdown
Please
1. run the tests with `npm test`
2. choose one substantive failing test
3. review @STATEVIS.md to see our instrumentation and debugging
4. integrate the instrumentions
5. run the individual test and collect the evidence from the debugging and instrumentation
6. perform any additional investigation, tracking the data through the system in order to eliminate any assumptions in understanding of the issue
7. create a new issue file in `dev/ISSUE-{something}.md` with the gathered evidence using the template in @ISSUE-TEMPLATE.md
# [ISSUE TITLE] Issue Investigation
## Issue Description
## Related Code
### Code Files
### Test Files
## Evidence
### Test Expectations
### Current Behavior
### Code Analysis
#### Test Structure
#### Related Tests
### Debug Output
#### Visualization
#### Instrumentation Analysis
## Attempted Solutions
## Additional Notes
## Next Steps