blueprintlisted
Install: claude install-skill rp1-run/rp1
# Project Blueprint
The top-level skill owns every user-facing question. The two retained agents are optional, bounded, non-interactive finalizers that run only after the parent's accepted answers are durable.
{% include_shared "parent-owned-interview.md" %}
## STATE-MACHINE
```mermaid
stateDiagram-v2
[*] --> detect
detect --> charter : charter_needs_work
detect --> prd : charter_complete
charter --> prd : charter_complete
prd --> [*] : prd_complete
```
On each phase transition, report with `rp1 agent-tools emit` using workflow `blueprint`, the generated `{RUN_ID}`, and the current state as `--step`. Derive `RUN_NAME` as `Blueprint: {EFFECTIVE_PRD_NAME}`. The first emit includes `--name "{RUN_NAME}"`. Enter a non-terminal state with `{"status":"running"}`; finish `prd` with `{"status":"completed"}` and `--close-run`. On a fatal phase error, emit `{"status":"failed"}` for the current state before stopping.
## Context
Use the pre-resolved `projectRoot`, `kbRoot`, and `workRoot` values from Workflow Bootstrap. They are authoritative and may point to different storage locations.
- Charter: `{kbRoot}/charter.md`
- PRD: `PRD_PATH`, derived after name validation as `{workRoot}/prds/{EFFECTIVE_PRD_NAME}.md`
- Charter template: `plugins/base/skills/artifact-templates/templates/charter-interviewer/charter.md`
- PRD template: `plugins/base/skills/artifact-templates/templates/blueprint-wizard/prd.md`
## Procedure
### 1. Validate The PRD Name Before Effects