backlog-discoverer

Discovers and triages GitHub issues to feed the pipeline with prioritised work items.

When to use

DISCOVER phase of the pipeline
When a user assigns an issue or a milestone
Automatic trigger via the orchestrator

Entry contract

Assigned GitHub issue or milestone identifier
Access to the GitHub repository (labels, issues, history)

Exit contract

Triage report with labels, priority, and effort estimate
Artifact written to the pipeline tracking directory

Invariants

Single-issue scope — Processes one issue at a time, never batched
Evidence-based triage — Every decision (label, priority, effort) is backed by observable data
See Customisation for the full list

Why this shape

The discoverer applies explicit review criteria for every triage decision. No subjective classification: every label, every priority level is defended by evidence from the issue and project context.

« Define explicit review criteria before the review begins. » — Wiegers, K., Peer Reviews in Software, 2002.

Single-issue scope (one issue = one cycle) prevents the batching that dilutes attention and produces superficial triages.

« The best XP teams treat scope, time, cost, and quality as variables to be consciously managed. » — Beck, K., Extreme Programming Explained, 2nd ed., 2004.

Allowed customisation

Priority criteria and label taxonomy (L1)
Analysis depth (quick vs thorough) (L2)
Triage report template (L1)