Generates roadmap from requirements by selecting walking skeleton iteration with first journey scenario, ordering stories into follow-on iterations, and splitting stories by AC dependency profiles.
npx claudepluginhub prime-radiant-inc/prime-radiant-marketplace --plugin iterative-developmentThis skill uses the workspace's default tool permissions.
Reads the per-epic requirement files in `docs/superpowers/iterations/requirements/` and `docs/superpowers/iterations/behavior-scenarios.md`, and produces `docs/superpowers/iterations/roadmap.md`: a walking-skeleton iteration (ITER-0000) plus ordered follow-on iterations. The walking skeleton must produce the first runnable journey scenario. Runs citation and scope review via PAR before committi...
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Searches prompts.chat for AI prompt templates by keyword or category, retrieves by ID with variable handling, and improves prompts via AI. Use for discovering or enhancing prompts.
Checks Next.js compilation errors using a running Turbopack dev server after code edits. Fixes actionable issues before reporting complete. Replaces `next build`.
Reads the per-epic requirement files in docs/superpowers/iterations/requirements/ and docs/superpowers/iterations/behavior-scenarios.md, and produces docs/superpowers/iterations/roadmap.md: a walking-skeleton iteration (ITER-0000) plus ordered follow-on iterations. The walking skeleton must produce the first runnable journey scenario. Runs citation and scope review via PAR before committing the roadmap.
Invoked by iterative-development during bootstrap after extracting-requirements.
All scripts referenced below live in this skill's scripts/ directory, next to this SKILL.md file.
Read the epic files in docs/superpowers/iterations/requirements/ — scan epic headers and story titles first, then dip into specific epic files for ACs when selecting.
Also read docs/superpowers/iterations/behavior-scenarios.md to understand which scenarios exist and which stories they cover.
Select a small cohesive set of stories from as many distinct epics as possible. The walking skeleton should prove the end-to-end shape of the product works.
Scenario requirement: the walking skeleton MUST include stories that close at least ONE journey scenario chain. Prefer the core product journey (e.g., "normal dictation" over "debug investigation"). The skeleton is not done until that journey scenario is runnable as an automated or scripted-reproducible test.
The walking skeleton must also produce:
Harness-first task: The walking skeleton's FIRST task should be designing and building the E2E test harness — before implementing any product features. Use the test infrastructure checklist in skills/shared/behavior-evidence-formats.md to guide the design. Document the harness design decisions in the project's docs. The harness is a first-class deliverable, not an afterthought — every subsequent iteration extends it.
Selection rule: "if someone ran just these stories, they should see a demo that proves the product exists AND have at least one passing journey scenario that proves the demo works."
Each iteration is a sprint's worth of cohesive work. Iteration granularity is judgment-based — no hardcoded story count.
Story splitting rule: when assigning stories to iterations, check each story's ACs for dependency profiles. If a story has ACs where:
Then SPLIT the story:
a/b to the story ID)Why this matters: moving a whole story to a later iteration because one AC has a late dependency causes the other ACs to be re-interpreted through the receiving iteration's theme, and they get silently dropped.
Run: python3 "scripts/check_citations.py" docs/superpowers/iterations/roadmap.md docs/superpowers/iterations/requirements/
Every iteration must cite only valid STORY-IDs from the index.
Following skills/shared/parallel-adversarial-review.md:
skills/running-an-iteration/scope-reviewer-prompt.mdWrite the result to docs/superpowers/iterations/roadmap.md using this format:
# Roadmap
## Walking skeleton (ITER-0000)
**Intent:** <one-line description of the thinnest end-to-end slice>
**Design rationale:** <why these stories, what they prove together>
**Journey scenario:** <JOURNEY-NNNN that the skeleton must pass>
**Stories committed:**
- STORY-NNNN (EPIC-NNN)
- ...
**Status:** pending
## Iteration list
### ITER-0001 — <name>
**Stories:** STORY-NNNN, STORY-NNNN, ...
**Rationale:** <why these stories belong together>
**Status:** pending
**Impacted scenarios:** <SCENARIO-NNNN, JOURNEY-NNNN that this iteration touches>
**Look-ahead check:** <does this block or get blocked by neighbors?>
Run: python3 "scripts/validate_roadmap.py" docs/superpowers/iterations/roadmap.md
Note: The validator checks format only. The PAR scope review is the real structural gate.
git add docs/superpowers/iterations/roadmap.md
git commit -m "docs: add roadmap — walking skeleton with journey scenario + iteration plan"
| Step | Tool/Skill | Purpose |
|---|---|---|
| Citation check | scripts/check_citations.py | All cited stories exist |
| Scope review | PAR + scope reviewer prompt | Walking skeleton minimal, journey scenario included, story splitting applied, no boxing-in |
| Story splitting | Manual (if PAR or dependency analysis finds heterogeneous ACs) | Split stories by dependency profile |
| Validate | scripts/validate_roadmap.py | Format check only |
skills/shared/parallel-adversarial-review.md — PAR methodologyskills/shared/behavior-evidence-formats.md — scenario and proof obligation formatsskills/running-an-iteration/scope-reviewer-prompt.md — scope reviewer prompt (reused)scripts/check_citations.py — mechanical citation check