Structured Thinking Template Packs

Use these packs when compact probes are not enough and you need a repeatable flow for a specific decision shape. See escalation criteria in structured-thinking-checklists.md.

Keep responses concise (usually 1-3 bullets per prompt) and attach outputs to existing artifacts (objective function, decision table, measurement ladder, finish packet).

Use template packs only for non-trivial work unless a tiny change has unusually high risk.

Quick Selector

Template	Use when	Primary skills
Technical Design Review	Architecture or boundary design choices need stress-testing	architecture, spec, review
Trade-Off / Project Decision	Multiple viable options must be compared with explicit opportunity cost	plan, spec, workflow
Retrospective / Postmortem	A delivery, incident, or rollout needs learning capture	finish, debug
Recommendation Brief	You must frame a PR/ADR/technical recommendation for async review	finish, review, workflow

See also: Appendix — Strategic Planning / Roadmap (edge case for multi-quarter work).

1) Technical Design Review

Use for non-trivial system/software design decisions before implementation locks in.

Prompts:

• Confirm fundamentals: what requirements/constraints are facts vs assumptions?
• Map system interactions: where are reinforcing/balancing feedback loops?
• Scan second-order effects: what likely changes now, next quarter, next year?
• Compare alternatives: what does each option optimize and what does it force us not to do?
• Risk + mitigation: what can fail, how will we detect/correct quickly?

Output (probe field mappings in parentheses):

• fundamentals (→ probe #1: facts + assumptions + assumption-to-test-first)
• interaction + loops (→ probe #3: reinforcing loop + balancing loop + delay + accumulation risk)
• second-order effects (→ probe #2: near-term effects + long-term effects + pre-mortem cause)
• alternatives + opportunity cost (→ probe #4: opportunity cost + bias risks)
• risks + mitigations
• recommendation

2) Trade-Off / Project Decision

Use when selecting one path among several project options.

Prompts:

• Define goals/constraints: what must be true for this decision to be valid?
• List options: include status quo/no-change baseline.
• Evaluate first-order pros/cons per option.
• Evaluate second-order effects ("and then what?" across time horizons).
• Check system impact: who absorbs downsides; which loops are triggered?
• Make opportunity cost explicit: what are we saying "no" to?
• Run a bias check (sunk cost/familiarity/novelty/confirmation).

Output (probe field mappings in parentheses):

• decision context (→ probe #1: facts + assumptions)
• option table
• short-term effects (→ probe #2: near-term effects)
• long-term effects (→ probe #2: long-term effects + deferred cost owner + pre-mortem cause)
• system + stakeholder impact (→ probe #3: reinforcing loop + balancing loop)
• opportunity cost + bias risks (→ probe #4: opportunity cost + bias risks + external challenge)
• selected option + rationale

3) Retrospective / Postmortem

Use after completion, incident, or rollback to convert outcomes into controls.

Prompts:

• Facts vs expectation: what happened, and what was expected?
• Cause/effect chain: what events produced the observed outcome?
• Root causes: what failed at the assumption/process/invariant level?
• Second-order effects: what quick fixes created delayed costs (or benefits)?
• Missed opportunities: what did we skip and what did it cost?
• Actions: what one process/control change would reduce repeat risk? (flag for human to assign owner + date)

Output:

• outcome delta
• cause chain
• root causes
• second-order effects
• missed opportunities
• owner-backed actions
• follow-up review date

4) Recommendation Brief

Use to frame a PR description, ADR recommendation, or technical proposal for async review. Keeps the recommendation grounded in evidence and actionable for reviewers.

Prompts:

• Core message: what action/decision is this PR/ADR requesting?
• Reviewer context: what do reviewers need to know to evaluate this (constraints, prior decisions, scope)?
• Evidence: what 2-3 facts or test results best support the recommendation?
• Counterpoints: what risks or alternatives remain and how are they mitigated?
• Open questions: what specific feedback is requested from reviewers?
• Next step: owner + date + what happens after approval.

Output:

• core recommendation
• reviewer context
• evidence
• counterpoints + mitigation
• open questions for reviewers
• owner + date + next step

Guardrails

• Prefer one template per decision checkpoint; do not run all packs.
• Keep outputs attached to existing artifacts, not separate long-form essays.
• If the work is tiny, skip both template packs and probes (see checklists).

---

Appendix — Strategic Planning / Roadmap

Edge case — use only for longer-horizon work (multi-quarter direction, initiative portfolio). Most agent-driven work fits the plan skill's normal implementation-planning focus; use this template only when the decision scope genuinely spans multiple teams or quarters.

Prompts:

• Vision + first principles: what core truths anchor this strategy?
• Current state + assumptions: which assumptions need validation or hedges?
• Scenario scan: if we do this, then what at 6/12/36 months?
• System map: what loops, delays, bottlenecks, and dependencies matter?
• Prioritization + opportunity cost: which initiatives make the cut and why?
• Pre-mortem: assume failure in 1-2 years; what likely caused it?
• Execution + feedback loops: what metrics/checkpoints trigger adaptation?

Output:

• vision + principles
• assumptions + validation plan
• scenario outcomes
• system dynamics notes
• priorities + explicit de-priorities
• pre-mortem risks + safeguards
• execution rhythm + checkpoints