Identify underspecified areas in the current feature spec by asking targeted clarification questions.
Detect and reduce ambiguity in the active feature spec. This should run BEFORE planning.
/clarify
Find the active spec:
git branch --show-currentspecs/in-progress/[branch-name]/specs/backlog/[branch-name]/specs/completed/[branch-name]/specs/[branch-name]/ (legacy fallback)spec.md from the resolved directory/specify firstLoad and scan the spec for ambiguities across these categories:
| Category | What to check |
|---|---|
| Functional Scope | Core goals, out-of-scope declarations |
| Data & Config | Env vars, Helm values, Terraform variables, types |
| Infrastructure | Kubernetes resources, Redpanda topics, NATS subjects |
| Performance | Connection limits, rate limits, resource requests/limits |
| Integration | Service interactions, failure modes, backpressure |
| Edge Cases | Crash recovery, rebalancing, pod preemption |
| Constraints | DOKS limits, node pool sizing, provider quotas |
Mark each: Clear / Partial / Missing
Generate questions (max 5 total):
Ask ONE question at a time:
After EACH accepted answer, update the spec:
## Clarifications section- Q: <question> → A: <answer>Update pass counter in spec.md:
**Passes**: line near the top of the spec (after the metadata block)clarify counter and update its status. Do NOT modify other skill counters.**Passes**: clarify: 1 ✓ (or without ✓ if issues were found) after the last metadata line (before ## Context)✓ if no ambiguities were found (clean pass), remove ✓ if issues were found/questions asked**Passes**: clarify: 2 ✓ means 2 passes, last one was cleanReport completion:
/plan-feature)