/plan-exposure
Generate an Exposure Plan, following the exposure-plans skill.
Input: $ARGUMENTS
Workflow
Resolve the source. A hypothesis file (
product/hypotheses/), a spec file (product/specs/), or a description given inline. If nothing resolvable, ask. If a hypothesis file exists for this feature, use its hypothesis verbatim as the spec-level hypothesis.Resolve the mechanism preference.
flagsorincrementalif passed — binding. Otherwise the skill recommends one and names the driver.Build the plan per the
exposure-plansskill: sequenced falsifiable beliefs, 1–5 accumulative user-visible levels with belief/audience/duration/validation, mechanism fit note, parallel validations, open questions. Same language as the source document.Present and confirm. Show the full plan in the conversation — never ask to save (or save) a plan the user hasn't seen. Walk them through the belief sequence and the levels; adjust until they own it.
Save to
product/exposure-plans/{YYYY-MM-DD-HHMM}-{slug}.md(timestamp = creation date; revise in place without renaming) once the user agrees.
Close in one line: suggest building Level 1 and wiring the instrumentation defined in the hypothesis before revealing anything.