ce-strategy

Creates and maintains STRATEGY.md, a short, durable anchor document that captures what a product is, who it serves, how it succeeds, and where the team is investing. The file lives at the repo root as a canonical peer of README.md, and downstream skills (ce-ideate, ce-brainstorm, and ce-plan) read it as grounding when it exists. Use it when starting a product, changing direction or roadmap, or when those downstream skills need upstream product grounding; an optional argument can focus the run on a section such as metrics or approach.

The skill runs a disciplined interview rather than free-form drafting. It defaults to the platform's blocking question tool (for example AskUserQuestion in Claude Code) and asks one question at a time, preferring free-form responses for substantive sections and single-select for routing decisions. Execution routes by file state: if STRATEGY.md does not exist it runs a first-run interview covering target problem, our approach, who it is for, key metrics, and tracks, plus optional milestones, not-working-on, and marketing sections; if the file exists it does a targeted or user-selected update, preserving sections that still hold. It reads references/interview.md for pushback rules and quality bars, applies up to two rounds of pushback per section, then fills references/strategy-template.md and presents a draft before writing, updating the last_updated frontmatter on updates.

Core principles keep the document an anchor rather than a plan: features belong in ce-brainstorm and schedules in the issue tracker, rigor lives in the interview questions rather than the headings, and short is treated as a feature so section expansion is pushed back on. The skill explicitly does not update the issue tracker, prioritize the backlog, write requirements or implementation plans, or compute metric values. Its structure is informed by Richard Rumelt's Good Strategy Bad Strategy, and after writing it hands off to the downstream ce-ideate, ce-brainstorm, and ce-plan skills.