Technical Debt Prioritizer

What it does

Technical-debt debates usually run on gut feel — "the ugliest file" or "the biggest god-class" — and reach the wrong conclusion. This skill replaces that instinct by lining up four accepted frameworks into a single, evidence-based priority order:

1. SQALE (Letouzey, ISO/IEC 25010 aligned) — assigns each item a remediation_cost (person-days) and interest_per_month (friction that accrues if unpaid), summarized as a Technical Debt Ratio = remediation / development cost.
2. WSJF / Cost of Delay (SAFe, Reinertsen) — WSJF = Cost of Delay / Job Size. For debt the heaviest Cost-of-Delay term is "risk reduction & opportunity enablement"; the tie-breaker is shortest job first.
3. Fowler's Technical Debt Quadrant — Reckless/Prudent × Deliberate/Inadvertent; intent and awareness feed the risk weight.
4. CodeScene-style Behavioral Code Analysis (Hotspot) — measures debt as change frequency × complexity, not static ugliness. Rarely-touched ugly code carries zero interest.

The synthesis lands in a TDPS (Technical Debt Priority Score, 0-100): (Risk × Interest × BlastRadius / RemediationEffort) × HotspotFactor × ConfidenceFactor, banded into Now (≥70) / Next (40-69) / Later (15-39) / Accept (<15), corrected by a dependency graph (blocked_by), with an estimate of lead-time recovered by paying the top items.

When to use it

• "Where is the debt in this codebase, and what do we fix first?"
• "Rank these 14 refactor candidates with reasoning."
• "We have 3 sprints in Q3 — which debt payoff has the highest return?"
• "Which module of this monolith is riskiest / most expensive?"
• When an architect, tech lead, staff engineer or EM must defend a refactor investment to leadership.

Method / frameworks

The score binds to named canon: SQALE (remediation/non-remediation cost, TD Ratio), WSJF & Cost of Delay (economic sequencing + shortest-job-first tie-break), Fowler's Quadrant (intent/awareness classification), Behavioral Code Analysis / CodeScene (hotspot = frequency × complexity), anchored to ISO/IEC 25010 maintainability sub-characteristics with interest measured from DORA signals. Industry anchor: SonarQube A-rating = TD ratio under 5%, healthy orgs stay below B; teams typically allocate 15-20% of each sprint to debt, and over 30% signals productivity drag. Where R/I/B/E come from measurement C=1.0, where estimated C≤0.8 + gaps — effort is given as a typical range, never a fake single number. It never recommends a big-bang rewrite; the Now band is always sprint-sized, reversible items.