Skip to content

Latest commit

 

History

History
113 lines (94 loc) · 5.61 KB

triage-issue.prompt.md

File metadata and controls

113 lines (94 loc) · 5.61 KB
name description argument-hint agent tools
triage-issue
Triage a new GitHub issue — categorize, label, request info, and link related items.
<issue number>
agent
github/search_issues
read/readFile
codebase/search

Triage GitHub issue #${input:issue} in dotnet/SqlClient.

GitHub Project: https://github.com/orgs/dotnet/projects/588/ — All issues are tracked here. After triaging, update the issue's project fields.

Follow this workflow step-by-step:

1. Fetch and Understand the Issue

  • Retrieve the full issue details (title, body, labels, author, comments).
  • Determine which issue template was used:
    • ISSUE_TEMPLATE/bug-report.md → Bug
    • ISSUE_TEMPLATE/feature_request.md → Feature
    • No template / sub-issue → Task
    • High-level work with sub-issues → Epic

2. Set Issue Type

Update the issue type if not already set:

  • Bug — Unexpected behavior, crash, regression
  • Feature — New capability or enhancement proposal
  • Task — Sub-issue or discrete work item
  • Epic — High-level work linking multiple sub-issues

3. Evaluate Completeness

Check whether the issue provides enough information to act on:

For Bugs, verify:

  • .NET version and target framework
  • Microsoft.Data.SqlClient version
  • SQL Server version (if relevant)
  • OS and platform
  • Repro steps or minimal repro code
  • Expected vs actual behavior
  • Stack trace or error message

For Features, verify:

  • Clear problem statement / motivation
  • Proposed solution or API shape
  • Use case / scenario description

If any required details are missing, add the label Needs more info :information_source: and post a comment requesting the specific missing items.

4. Apply Labels

Select appropriate labels based on the issue content:

Area labels (pick the most relevant):

  • Area\Engineering — Build system, CI/CD, project infrastructure
  • Area\Connection Pooling — Pool behavior, timeouts, pool size
  • Area\AKV Provider — Always Encrypted Azure Key Vault provider
  • Area\Json — JSON data type support
  • Area\SqlClient — General driver behavior
  • Area\SNI — Network layer (managed or native)
  • Area\TDS — Protocol-level issues

Status labels:

  • Triage Needed :new: — For new issues needing initial review
  • Needs more info :information_source: — Missing required details
  • Performance :chart_with_upwards_trend: — Performance-related concern

5. Update GitHub Project Fields

Ensure the issue is added to the GitHub Project (https://github.com/orgs/dotnet/projects/588/) and set the following fields to the most appropriate values:

Field Values Guidance
Status To Triage, Needs Response, Investigating, Waiting for customer, Backlog, In progress, In review, Done Set to Investigating if actionable, Needs Response if info is missing from reporter, Waiting for customer if awaiting external input, Backlog for valid lower-priority items
Priority P0, P1, P2, P3 P0: Critical — data loss, security vulnerability, widespread breakage. P1: High — significant regression, blocking customer scenario. P2: Medium — important but has workaround, affects subset of users. P3: Low — minor issue, enhancement, nice-to-have
Size XS, S, M, L, XL Estimate effort: XS: trivial fix (<1h). S: small, well-scoped (<1 day). M: moderate, may touch multiple files (1-3 days). L: significant, cross-cutting (1-2 weeks). XL: large feature or architectural change (2+ weeks)
Comment Free text Add a brief note explaining priority/size rationale or any context for the team

Field Assignment Examples

  • Bug with repro + clear root cause → Status: Investigating, Priority: P2, Size: S
  • Bug with missing info → Status: Needs Response, Priority: unset until understood
  • Feature request, well-specified → Status: Backlog, Priority: P3, Size: M
  • Security vulnerability → Status: Investigating, Priority: P0, Size: varies
  • Performance regression → Status: Investigating, Priority: P1, Size: M

6. Search for Related Issues

  • Search for existing issues with similar keywords: repo:dotnet/SqlClient <key terms>
  • If duplicates or related issues exist:
    • Link them in a comment
    • If it's an exact duplicate, label accordingly and reference the original
  • Check if there are related PRs (open or merged) that address the same area.

7. Identify Affected Code Area

  • Based on the issue description, identify the likely source file(s) in src/Microsoft.Data.SqlClient/src/Microsoft/Data/SqlClient/.
  • Note the component (e.g., SqlConnection, TdsParser, ConnectionPool, SqlCommand).
  • Check if the issue is platform-specific (Windows-only, Unix-only, .NET Framework-only).

8. Post Triage Comment

Add a comment that includes:

  • Confirmation that the issue has been triaged
  • Issue type classification
  • Which area/component is likely affected
  • Any questions or requests for missing information
  • Links to related issues, documentation, or code references
  • GitHub Project fields that were set (Status, Priority, Size)
  • @ mentions for relevant team members if the issue is urgent or high-impact

9. Summary

Output a brief triage summary:

  • Type: Bug / Feature / Task / Epic
  • Area: Which component(s) affected
  • Labels applied: List of labels
  • Project fields: Status, Priority, Size values set
  • Missing info: What additional info is needed (if any)
  • Related issues: Links to related issues or PRs
  • Severity assessment: Low / Medium / High / Critical