| name | strategy |
| description | Replan an existing product or project after context changes. |
| argument-hint | project context, pivot, launch, roadmap, or constraints |
| license | MIT |
| metadata | {"category":"strategy","tags":["strategy","roadmap","backlog","adr"],"effort":"high","side_effect":"docs-write","explicit_invocation":false} |
Strategy
Update direction for an existing project.
The canonical strategy rubric lives in docs/rubrics/strategy.md.
Outputs
As needed:
MARKET_BRIEF.md
ICP.md
BUSINESS_MODEL.md
GO_TO_MARKET.md
IDEA_SCORECARD.md
VALIDATION_PLAN.md
FUNNEL_MODEL.md
FEEDBACK_LOG.md
ITERATION_REVIEW.md
.pfo/EXPERIMENT_PROGRAM.md
.pfo/EXPERIMENTS.tsv
ASSET_REGISTER.md
CONTENT_BACKLOG.md
LAUNCH_PLAN.md
LAUNCH_MATURITY_GATE.md when launch-stage ops readiness is in scope
SCALE_MOAT_REGISTER.md when scale, enterprise, or defensibility is in scope
BACKLOG.md
docs/adr/YYYY-MM-DD-decision.md
Process
- Read current docs and project state.
- Apply
docs/rubrics/strategy.md.
- Run
/market-scan when roadmap, pivot, launch, ICP, competitor, pricing, or feedback decisions depend on fresh public market/community signals.
- Identify gaps across product, tech, delivery, operations, monetization, acquisition, and go-to-market.
- Define or update:
- problem
- ICP
- value proposition
- MVP scope
- idea score and KILL/TEST/BUILD decision
- evidence quality: real conversations, past behavior evidence, contradicting evidence, BUILD truth conditions
- validation experiments and exit criteria
- adversarial discovery: failure case, competitor win case, current alternatives, ignored uncomfortable signals
- monetization or value-capture model
- acquisition path
- funnel stages, MVP measurement contract, false-positive traction, and bottleneck
- feedback loop and iteration decision
- optional launch maturity audit when the founder or operator is becoming a bottleneck
- optional scale moat register when domain knowledge, edge cases, data flywheel, workflow integration, or switching cost matter
- fixed metric experiment loop when autonomous self-improvement is requested
- asset and content candidates
- launch success metrics
- strategic risk matrix
- Generate options and tradeoffs.
- Record the chosen direction as an ADR.
- Update backlog and launch plan.
- Update
.codex-memory/STATE.json decision log when present.
Self-validation
Before final output, verify:
- Route, side-effect, and confirmation requirements match metadata.
- Required artifacts or read-only result are explicit.
- Verification, blockers, and next route are stated.
Rules
- Planning docs only unless the user asks for implementation.
- Do not rewrite history; append dated decisions.
- Do not allow large implementation work to start when strategy Critical checks fail for a non-trivial product.
- Do not count activity as progress unless it changes a signal, gate, decision, or reusable asset.
- Do not count founder enthusiasm, compliments, or future-intent answers as strong demand evidence.
- For Autoresearch-style work, use one metric, fixed budget, protected evaluation, baseline-first logging, and keep/discard/crash decisions.
- Do not treat recent social buzz as validated demand. Convert it into
VALIDATION_PLAN.md experiments and decision criteria.