issue-triage

Assigns structured metadata to GitHub issues (type, priority, effort, duplicates) to enable the DISCUSS phase to refine them in priority order.

When to use

Classifying newly created GitHub issues (status/needs-triage)
Assigning type, priority, and effort labels to an issue
Detecting duplicates and related issues
Building a sprint proposal with calculated capacity
Flagging XL issues that must be split before entering DISCUSS

Entry contract

List of GitHub issues with title and description
Access to the mcp_github_issue_write MCP tool to apply labels

Exit contract

Triage table with: number, title, type, priority, effort, notes
Labels applied to each issue: type/*, priority/*, effort/*, status/ready or status/duplicate
Sprint proposal with calculated capacity
“Must split” notes for every effort/XL issue

Invariants

Triage is classification, not refinement — Acceptance criteria are not written here
Every triaged issue must have a type, a priority, and an effort before status/ready
P0 requires written justification — Notes field mandatory: reason, impact, date
XL must be flagged — Any effort/XL issue blocks entry to DISCUSS without a splitting plan
Fixed label taxonomy — type/feature, type/bug, type/tech-debt, type/docs, type/question for types; priority/P0–P3 for priority; effort/XS–XL for effort

Why this shape

Triage establishes a classification discipline before any refinement work. Separating classification (DISCOVER) from refinement (DISCUSS) avoids wasting energy refining low-priority issues or duplicates. The priority decision tree (P0 → P1 → P2 → P3) guarantees consistency across teams.

« Triage is classification, not refinement. You are labeling, prioritizing, and estimating to enable the next phase. »

Duplicate detection via title normalisation (lowercase, stop-word removal, alphabetical sort, overlap ratio) reduces backlog noise without losing the context of original issues.

Allowed customisation

Area labels (area/*) (L1)
Similarity thresholds for duplicate detection (L2)
Effort-to-days conversion for sprint capacity (L2)