Roadmap
Prompts don't compound. Skills do. The open-source AI toolkit for product managers — 13 Claude Code skills + 3 red-team agents for the full PM workflow.
npx -y skills add ramanbamba/10x-pm --skill roadmapAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Build a Now/Next/Later product roadmap organized by themes with confidence levels and kill criteria. Use when the user says "build a roadmap", "roadmap for next quarter/half/year", "now next later", or needs to turn a strategy and backlog into a communicable plan.
SKILL.md
3.2 KB, as published. Nobody here has run it
Roadmap
Build a roadmap that communicates intent and confidence, not fake dates. Now/Next/Later by default; date-committed only where a real external commitment exists.
Before building
- Ask for the strategy or goal the roadmap serves. A roadmap is strategy made visible — without one, it's a feature list wearing a suit. If no strategy exists, offer to draft a one-paragraph strategy statement first and get it confirmed.
- Ask who the audience is: internal team, execs, sales, or public. The same roadmap has different altitude for each — confirm before formatting.
- Ask which items carry hard external commitments (contracts, compliance dates, announced features). Only these get dates.
Workflow
- Organize by theme (problem/outcome), not feature. "Reduce time-to-first-value" is a theme; "onboarding wizard" is one bet inside it. Themes survive when individual bets die.
- Place bets into Now / Next / Later:
- Now = in progress or starting within weeks. High confidence (≥80%). Named teams.
- Next = after Now, roughly this quarter/next. Medium confidence — shaped but re-orderable.
- Later = directional. Low confidence, explicitly so. No item counts or sizes here; it's a statement of interest, not a plan.
- Attach to every Now/Next bet: the theme it serves, the outcome metric it should move, confidence %, and a kill criterion — the observable result that would make you stop or pivot the bet.
- Write the "not doing" list. 3–5 plausible items deliberately excluded, with one-line reasons. This is the highest-signal section for stakeholders.
- State the review cadence and what would trigger an off-cycle change.
Output format
# Roadmap — [product], [horizon], as of [date]
**Strategy:** [one paragraph this roadmap serves]
**Audience:** [who] · **Review cadence:** [when]
## Now
### Theme: [outcome-framed theme]
| Bet | Outcome metric it moves | Confidence | Kill criterion | Team |
|---|---|---|---|---|
## Next
[Same structure, medium confidence, no teams required]
## Later
[Themes and one-line bet descriptions only. Header note: "Directional — sequencing and scope will change."]
## Committed dates
[ONLY externally-committed items: item · date · nature of commitment]
## Not doing (deliberately)
| Item | Why not |
|---|---|
## Change log
[What moved since last version and why — starts empty]
Quality bar — self-check
- Every bet maps to a theme and a metric. Orphan features get questioned or cut.
- Confidence is explicit and Later is visibly uncertain — if Later looks as firm as Now, the roadmap over-promises.
- Every Now/Next bet has a kill criterion. Roadmaps without kill criteria are ratchets: things get on, nothing gets off.
- Dates appear only with real external commitments. Internal targets are confidence levels, not dates.
- The "not doing" list is non-empty and specific.