What this skill IS

This is strategic problem selection, not issue triage. You are managing a portfolio of problems — deciding which problems to invest effort in, balancing exploration vs exploitation, maintaining diversity, and preserving stepping stones.

Key questions while filling the template:

• Which problems are in the zone of proximal development (goldilocks)?

• Am I over-investing in one type of problem (exploit bias)?

• Which problems, even if not urgent, could open new solution spaces (stepping stones)?

• What problem types are NOT in the portfolio but should be?

Prerequisites

• ≥2 PROB/ANOM-* cards exist (otherwise no portfolio needed)

• If comparing problems on measurable dimensions: invoke /fpf-characterize for problem-level characterization. Q dimensions in the portfolio table MUST reference CHR-* indicators, not hardcoded defaults.

Output

.fpf/portfolios/PPORT-${CLAUDE_SESSION_ID}--<slug>.md

Constraints (quality bar)

• C1: Selection rule stated BEFORE applying — not post-hoc rationalized

• C2: Archive complete — no hidden problems

• C3: ≥1 active problem has stepping-stone potential (not all exploit)

• C4: Diversification check honest — flags imbalances (problem type, domain, risk level)

• C5: Each active problem references existing PROB/ANOM-*

• C6: Goldilocks filter applied — trivial and impossible problems identified and deferred

• C7: Q dimensions in NQD table reference CHR-* indicators when characterization exists (not hardcoded impact/feasibility)

Selection policy examples

• Impact × Feasibility: highest expected value first (exploit-heavy)

• Learning value: problems that teach the most about the domain (explore-heavy)

• Barbell: 80% safe + 20% speculative stepping stones

• Constraint-driven: solve blocking problems first, then expand

Format

Problem Portfolio

• ID: PPORT-... Scope: ... E/E policy: explore|exploit|barbell

Selection rule

(policy stated before applied, chosen from above or custom)

Active problems (NQD-characterized)

| # | ID | Title | D_c | Q:[impact] | Q:[feasibility] | Q:[stepping-stone] | N | D_p | Pipeline: D_c (goldilocks filter) → Q dominance (Pareto) → N, D_p (tie-breakers). Never collapse Q to single score.

Deferred

| # | ID | Title | Why deferred | Revisit condition |

Diversification check

• Problem types covered: ...
• Not covered: ... (flag if imbalanced)
• Explore/exploit balance: ...