microsoft / microsoft/GitHub-Copilot-for-Azure
[Epic] Triage and resolve skill and deployment regressions continuously
- Dominant language
- Python
- Stars
- 250
- Forks
- 204
- Avg merge
- 1d 12h
- Merged PRs (30d)
- 67
Description
## Problem statement
Skill invocation, tool selection, generated guidance, deployment, and integration-test regressions accumulate as the product and its dependencies change. They need an operational path distinct from strategic quality investment.
## Outcome
Regressions in shipped skills and deployment workflows are reproduced, prioritized, fixed, and protected by durable tests.
## Goals
- Triage current product and integration-test failures
- Identify canonical issues for recurring failure patterns
- Add targeted regression coverage with each fix
- Track remediation through release validation
## Non-goals
- Using this Epic for new capabilities
- Retaining duplicate or non-actionable failure reports indefinitely
## Success criteria
- [ ] Active regressions have owners and reproduction evidence
- [ ] Duplicate failures converge on canonical issues
- [ ] Fixes include appropriate tests or evaluations
## Dependencies
Integration runners, skill owners, Azure MCP, deployment tooling, and release validation.
Contributor guide
Research direction
Review the active skill, deployment, and integration-test failures first, using the integration runners, Azure MCP, deployment tooling, and release validation named in the issue. Identify canonical issues and reproduction evidence, then ensure each remediation has targeted regression coverage and release-validation tracking.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- azure, python
- Domain
- cloud, devops, testing-qa, tooling
- Issue type
- Bug
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Quiet
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100