Creative Problem Solver

Purpose: when a strategy choice is still open, transform the incumbent into materially different candidate paths, project the honest survivors into a five-tier portfolio, and stop for a human choice.

Activation boundary

Implicit invocation is enabled. Host loading is not portfolio authorization. The portfolio route requires an unresolved strategy choice plus a request for divergent options. Difficulty, uncertainty, creativity language, or a name-only or meta mention is not enough.

Activate when

• The user asks for options, alternatives, trade-offs, fresh angles, reframing, a strategy portfolio, or "what else could we try?"
• Repeated attempts fail in the same way and selecting a materially different approach is the next task.
• A multi-constraint decision would benefit from several distinct conceptual frames before commitment.
• The primary deliverable is a choice set of materially different paths followed by human selection.
• A direction is selected but materially different rollout, migration, containment, or proof strategies remain open.

Do not activate; route elsewhere

• Direct implementation, debugging, review, or execution after an approach is selected: use the owning workflow.
• Mixed requests where options are subordinate to "choose and implement now": the execution owner may use this reasoning internally, but this skill must not seize the turn and discard the requested execution.
• A factual answer, ordinary creative generation such as names or copy, or comparison of a small known set where a portfolio adds ceremony: answer directly.
• Evidence-backed repository or product opportunity mining, ranked improvement discovery, or choosing what to plan next: use $ideate.
• Turning a selected direction into a detailed implementation plan, specification, or work graph: use $plan or $spec-pipeline.
• Architecture or codebase understanding without a request for divergent paths: use direct analysis or $codebase-archaeology.
• Analysis, explanation, review, tuning, testing, or editing of this skill itself: use direct analysis, $tune, or $refine as appropriate.
• The user already chose a tier or path and asks to execute it: hand off to the execution owner; do not regenerate the portfolio.

Skill-name rule

• Imperative requests to "use", "run", or "apply $creative-problem-solver" to generate a strategy portfolio explicitly authorize the portfolio route.
• A name-only or meta mention may load the package under host policy, but it routes to the task owner—or to direct clarification when no task exists—and does not authorize portfolio generation.

Tie-breakers

• Explicit imperative portfolio invocation wins unless safety, a domain-specific owner, a meta-task owner, or a contradictory execution request requires another route.
• $ideate wins when repository evidence mining and ranked opportunities are central.
• $plan wins when the direction is selected and the requested output is an execution policy or detailed plan.
• $tune / $refine win when the object of work is the skill package itself.
• Activate when the requested outcome is a choice set; do not activate when the outcome is one answer, one plan, or one implementation.

Method

1. Bind only what governs the choice:

   Objective: <outcome or decision to improve>
   Hard constraints: <non-negotiable boundaries>
   Incumbent/default basin: <the approach or representation most likely to be repeated>
   Material unknowns: <unknowns that could change the choice>
   

2. Diverge before formalizing: transform the incumbent across several orthogonal operator axes. Do not use the five tiers as idea prompts.
3. Quotient: collapse candidates that share a governing mechanism and differ only by cost, scope, implementation size, rollout speed, timeline, or ambition.
4. Aha Check after the initial field exists. If a material re-representation changes the field, proof surface, ordering, or decision criteria, regenerate once under it and quotient again. Otherwise report Aha: N/A without penalty.
5. Winnow pairwise non-subsumed candidates that would lead to meaningfully different commitments.
6. Validate proportionally: attach the smallest reversible probe, decision discriminator, kill criterion, retained asset, and pre-irreversible escape hatch appropriate to each survivor.
7. Project survivors onto Quick Win through Moonshot without changing their governing mechanisms to fit the labels. Preserve every heading, but never fabricate a candidate merely to fill a tier.
8. Recommend conditionally, then stop for human input before execution.

An unknown blocks solution candidates only when it could invert the decision, makes comparison meaningless, crosses a serious safety/legal/irreversibility boundary, or leaves no concrete decision surface. Otherwise state the assumption, branch conditionally, or resolve it through a probe. Never silently choose a domain, metric, threshold, denominator, population, repetition count, causal mechanism, or factual basis. Separate decision-shaping facts, supported inferences, assumptions, and hypotheses; label only those that affect the choice.

Lanes

• Fast Spark: default. Perform a hidden micro-divergence pass across at least four operator axes, including a subtractive move and a boundary or representation shift when admissible. Show only the decision-useful survivors.
• Full Session: use when the user explicitly requests deep or exhaustive exploration, or when high stakes, irreversibility, or coupled constraints make the candidate field itself decision-relevant. Generate roughly 10-30 raw candidates, cluster by governing mechanism, and expose only the clusters or transformations that changed winnowing.
• Difficulty alone does not justify Full Session. Visible brevity never authorizes shallow search.

Divergence kernel

Treat these as search transformations, not headings to fill:

• Subtract — eliminate the need, component, coordination, handoff, or inherited assumption.
• Invert — reverse a governing assumption, default, control direction, dependency, or value exchange.
• Move the boundary — change the owner, authority, interface, trust boundary, or substrate.
• Re-represent — change the model, data shape, abstraction, unit of work, or problem statement.
• Change time — precompute, defer, batch, stream, amortize, speculate, or make work event-driven.
• Change the actor or incentive — alter who acts, who benefits, who pays, or what behavior the system rewards.
• Change the proof surface — solve for a different observable, contract, falsifier, or acceptance boundary.
• Split, compose, or transfer — decompose the problem, combine separate moves, or import a mechanism from a structurally analogous domain.

Honor a user-requested creative technique when usable, but use it to transform the incumbent rather than as visible process theater.

Material diversity

• Selecting one survivor instead of another must change at least one of: governing mechanism, owner, representation, authority boundary, admitted domain, unit of work, proof obligation, or falsifier.
• Retain at least one subtractive candidate and at least one boundary or representation shift when those moves are admissible.
• A candidate subsumed by a stronger candidate is not separate unless it creates a materially different reversible entry point.
• If honest divergence is unavailable, say so. A same-basin portfolio or an empty tier is preferable to fabricated novelty.

Aha Check

• Aha is the restructuring insight: the problem is re-represented so a different candidate field becomes visible.
• Output Aha: <restructuring insight> or Aha: N/A.
• Follow a material Aha with Aha basis: <why the shift is warranted> [evidence: fact | supported inference].
• Put any empirical proposition on a separate optional line: Claim: <claim> [evidence: fact | supported inference | hypothesis].
• A material Aha must introduce a previously unavailable candidate, change a discriminator or proof surface, or reorder the surviving field. Regenerate once under it; otherwise it is commentary, not an Aha.
• Creative force never upgrades a basis or claim's factual status. A hypothesis needs a concrete disconfirming observation in a relevant probe.

Probes, evidence, and accretion

Every populated option must contain:

• Probe — the smallest reversible commitment that can change the decision; include a timebox when it materially improves control.
• Discriminator — the observation or comparison that would favor this path over its live alternatives.
• Kill criterion — the observation that defeats the path, its governing claim, or its admissibility.
• Retained asset — durable evidence, data, interface, prototype, test, model, specification, or decision record kept even when the path loses.
• Escape hatch — the stop, flag, isolation boundary, or pre-commitment that exists before irreversible authority, data, compatibility, or organizational commitment crosses the boundary.

Strengthen only as the claim requires: quantitative comparisons need whatever baseline, denominator, comparison rule, threshold, and repetition make the result interpretable; qualitative inquiry needs an evidence source, interpretation rule, and disconfirming evidence; causal hypotheses need falsifiers; irreversible commitments need abort boundaries and consequence containment. Do not force benchmark grammar onto qualitative or strategic decisions.

An Evidence Spine may hold neutral facts, baselines, examples, criteria, or evaluation harnesses that compare candidates without prescribing their construction. Each candidate may instead retain its own Option Artifact when its representation, interface, owner, or proof surface differs. Never force a transformative path to inherit the Quick Win's construction merely to share an artifact. Completion alone is not proof of benefit, and post-hoc quarantine is not reversibility.

Tier projection

• Quick Win — cheapest reversible move that improves the decision or outcome now.
• Strategic Play — strongest path within the current operating model.
• Advantage Play — creates a reusable capability, compounding asset, or asymmetric option.
• Transformative Move — changes an interface, operating model, governing constraint, authority boundary, or user ritual.
• Moonshot — tests discontinuous upside through the smallest bounded proof-bearing probe.

Project after winnowing. Do not escalate one mechanism through five sizes. If no honest candidate fits a tier, write No honest <tier> survived divergence and explain the missing distinction in one sentence.

Option card

<Tier> — <path> [shift: <mechanism, boundary, representation, or proof surface>]
- Probe: <smallest reversible test; timebox when meaningful>
- Decide by: <discriminator>; kill if <disconfirming observation>
- Retain: <durable option artifact>; escape before <irreversible boundary>

Add an assumption or evidence label only when it changes how the option should be interpreted.

Selection and output

• State the condition under which each leading path dominates and give one conditional recommendation.
• Prefer qualitative trade-off dominance over pseudo-precision. Use numeric scoring or Pareto language only when the user supplies criteria, scale anchors, and evidence that make the values meaningful.
• Fast Spark shows the compact binding, Aha or honest N/A, five tier headings with concise cards or honest gaps, one recommendation, and Human Input Required: choose a path or update the constraints.
• Full Session adds a compact candidate-field summary: mechanisms explored, equivalence classes collapsed, and transformations that changed winnowing.
• Do not expose the full operator checklist or discarded idea dump unless the user asks.
• A recommendation is not execution authorization. After selection, hand detailed planning to $plan and authorized implementation to the owning workflow. This skill never owns repository mutation.

Routing examples

Should activate

• "Give me several materially different ways to reduce onboarding drop-off before we choose one."
• "We have tried tuning the query twice and are still stuck. What else could we try?"
• "Reframe this migration problem and show the trade-offs among distinct paths."
• "The problem is ambiguous; give me learning moves rather than a build plan."
• "We selected event sourcing; show materially different rollout and containment strategies before execution."

Should not activate

• "$creative-problem-solver" -> direct clarification; no portfolio object is authorized.
• "Use $creative-problem-solver." -> direct clarification until a portfolio object is supplied.
• "Do a deep analysis of my $creative-problem-solver skill." -> direct analysis / $tune
• "Patch $creative-problem-solver so it stops over-triggering." -> $tune / $refine
• "I am blocked by this compiler error. What else can I try?" -> debugging owner
• "Brainstorm twenty names for this command." -> direct creative generation
• "Give me options, choose the best, and implement it now." -> execution owner; portfolio may be internal
• "Mine this repository for evidence-backed product and DX opportunities." -> $ideate
• "Turn the chosen event-sourcing direction into a detailed implementation plan." -> $plan
• "Explain this repository's architecture and data flow." -> direct analysis / $codebase-archaeology
• "Review this pull request for defects." -> review owner
• "We chose the Quick Win. Build it now." -> execution owner