Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Same roadmap, three audiences. Produce executive, customer, and internal roadmap variants with audience-appropriate framing, detail, and risk language. Use for quarterly roadmap publication, board packets, and customer/sales roadmaps.
.claude/skills/borghei-roadmap-communication/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 79% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -3% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 65% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 124% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 116% | 0% |
A single roadmap cannot serve every audience. Executives need confidence and strategic framing. Customers need plausible promises with credible timelines. Internal teams need detail, dependency awareness, and honest risk. When PMs publish one roadmap to all three, one audience is always wrong -- either the executives are bored, customers feel misled, or engineers feel managed.
This skill produces three variants of the same underlying roadmap, each tailored to its audience. The underlying source of truth -- the roadmap data -- stays single; the framing, level of detail, risk language, and visualization all change. The variants share names and identifiers so a customer reference can be traced back through the internal variant to the actual ticket. The structural backbone is Now / Next / Later (Janna Bastow, ProdPad), which Marty Cagan calls "right-sizing the roadmap" -- forecast confidence decreases with time horizon, so commit explicitly to "Now," name "Next" by theme, and keep "Later" deliberately fuzzy.
../scrum-master/ outputs).status-update-generator/).Before generating the variants, confirm these inputs. If any is unknown or vague, ASK — do not assume:
Stop rule: ask only the 2-3 that most change the output. If the user says "just draft it," proceed and list your assumptions at the top of the artifact.
assets/internal_roadmap_template.md.assets/executive_roadmap_template.md.assets/customer_roadmap_template.md.Load the reference that matches the task — keep this file lean and pull detail on demand:
In Scope:
Out of Scope:
prioritization-frameworks/)brainstorm-okrs/)north-star-metric/)outcome-roadmap/)../scrum-master/)dependency-map/)Important Caveats:
| Integration | Direction | Description | |-------------|-----------|-------------| | execution/outcome-roadmap/ | Receives from | Outcome decomposition feeds the executive variant | | execution/north-star-metric/ | Receives from | NSM and inputs are the executive variant's metrics | | execution/brainstorm-okrs/ | Receives from | KRs map to internal roadmap items | | execution/create-prd/ | Receives from | Each Now-quarter feature in the internal variant links to its PRD | | execution/prfaq/ | Receives from | PR/FAQ summaries become the seed of customer variant theme descriptions | | execution/prioritization-frameworks/ | Receives from | Prioritization scores determine which items make Now / Next / Later | | execution/dependency-map/ | Pairs with | Internal-variant dependencies are sourced from the dependency map | | execution/status-update-generator/ | Feeds into | Weekly status updates reference roadmap variant deltas | | ../senior-pm/ | Feeds into | Portfolio roll-ups aggregate executive variants across teams | | marketing/ (skills domain) | Hands off to | Customer variant feeds product marketing content calendars and launch comms |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 11,409 | 9,751 | -15% | 1 | 1 | 0% | 1,827 | 3,265 | +79% | 0 | 0 | — |
case-02 | fail→pass | 19,799 | 8,350 | -58% | 1 | 1 | 0% | 3,197 | 3,093 | -3% | 0 | 0 | — |
case-03 | pass→pass | 21,498 | 22,709 | +6% | 1 | 1 | 0% | 3,167 | 5,217 | +65% | 0 | 0 | — |
case-04 | fail→pass | 17,247 | 18,624 | +8% | 1 | 1 | 0% | 2,749 | 4,532 | +65% | 0 | 0 | — |
case-05 | fail→pass | 10,910 | 13,644 | +25% | 1 | 1 | 0% | 1,667 | 3,740 | +124% | 0 | 0 | — |
case-06 | pass→pass | 16,784 | 14,844 | -12% | 1 | 1 | 0% | 2,531 | 3,907 | +54% | 0 | 0 | — |
case-07 | fail→pass | 14,711 | 19,545 | +33% | 1 | 1 | 0% | 2,149 | 4,632 | +116% | 0 | 0 | — |
case-08 | fail→pass | 18,191 | 22,510 | +24% | 1 | 1 | 0% | 2,693 | 5,131 | +91% | 0 | 0 | — |
case-09 | pass→pass | 13,759 | 16,731 | +22% | 1 | 1 | 0% | 2,029 | 4,210 | +107% | 0 | 0 | — |
case-10 | fail→fail | 13,290 | 22,727 | +71% | 1 | 1 | 0% | 2,113 | 5,565 | +163% | 0 | 0 | — |
case-11 | pass→pass | 12,323 | 8,984 | -27% | 1 | 1 | 0% | 1,832 | 2,993 | +63% | 0 | 0 | — |
case-12 | pass→pass | 12,231 | 13,189 | +8% | 1 | 1 | 0% | 1,792 | 3,649 | +104% | 0 | 0 | — |
case-18 | fail→fail | 15,375 | 11,060 | -28% | 1 | 1 | 0% | 2,173 | 3,245 | +49% | 0 | 0 | — |
case-13 | pass→pass | 10,704 | 6,321 | -41% | 1 | 1 | 0% | 1,472 | 2,584 | +76% | 0 | 0 | — |
case-14 | fail→pass | 10,796 | 20,390 | +89% | 1 | 1 | 0% | 1,725 | 4,986 | +189% | 0 | 0 | — |
case-15 | pass→pass | 11,971 | 14,041 | +17% | 1 | 1 | 0% | 1,809 | 3,737 | +107% | 0 | 0 | — |
case-16 | pass→pass | 13,177 | 14,214 | +8% | 1 | 1 | 0% | 1,994 | 3,782 | +90% | 0 | 0 | — |
case-17 | pass→pass | 10,743 | 11,445 | +7% | 1 | 1 | 0% | 1,657 | 3,376 | +104% | 0 | 0 | — |
case-19 | pass→pass | 16,099 | 13,755 | -15% | 1 | 1 | 0% | 2,335 | 3,697 | +58% | 0 | 0 | — |
case-20 | fail→pass | 19,643 | 4,208 | -79% | 1 | 1 | 0% | 3,177 | 2,388 | -25% | 0 | 0 | — |
case-21 | fail→fail | 12,675 | 10,754 | -15% | 1 | 1 | 0% | 2,201 | 3,417 | +55% | 0 | 0 | — |
case-22 | fail→pass | 19,435 | 5,494 | -72% | 1 | 1 | 0% | 2,782 | 2,508 | -10% | 0 | 0 | — |
case-23 | fail→fail | 15,816 | 15,412 | -3% | 1 | 1 | 0% | 2,770 | 4,022 | +45% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted. The headline lift of +39 percentage points is the difference between those two pass rates over the 23 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.