Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Productboard expert for workspace setup, Insight-to-Feature triage, Driver scoring, Releases, Roadmap views, and REST API automation. Use for Productboard administration, prioritization workflows, and programmatic API operations.
.claude/skills/borghei-productboard-expert/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 5% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 47% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 60% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 77% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 204% | 0% |
Master-level expertise in Productboard workspace configuration, Insight inbox triage, Driver-based prioritization, Feature hierarchy management, Releases and Roadmap views, REST API operations, and two-way integration with Jira, Linear, Azure DevOps, Salesforce, Intercom, and Slack. Productboard sits between customer-feedback intake (Slack/Intercom/Salesforce/email) and the engineering tracker (Jira/Linear), with an opinionated three-layer model that differs meaningfully from generic issue trackers.
Productboard separates three layers: Insights (inbound customer evidence), Features (candidate product changes organized in Components and parent/child hierarchy), and Drivers + Releases (the prioritization framework — Drivers = weighted criteria, Releases = time-boxed delivery groupings). Insights link to Features (many-to-many); Features score against Drivers; high-scoring Features get added to Releases; Releases push down to Jira/Linear epics. The flow is inbox → triage → prioritize → ship.
The job of a Productboard expert is to operate fluently inside this model: configure the workspace cleanly, run the Insight inbox without backlog, design Driver weightings that match company strategy, and write API automations for what the UI does not cover.
Load the reference that matches the task — keep this file lean and pull detail on demand:
assets/productboard-feature-template.md (Feature structure), assets/productboard-insight-triage-workflow.md (daily triage SOP), assets/productboard-driver-template.md (Driver definitions). Official docs: https://developer.productboard.com/ · https://help.productboard.com/In Scope: Workspace and Component hierarchy setup, Insight inbox triage, Driver configuration and Feature scoring, Releases and Roadmap views (internal and customer-facing), Productboard REST API operations (Features, Insights, Releases, custom fields, webhooks), two-way integration setup with Jira/Linear/Azure DevOps, customer-data sync from Salesforce/HubSpot, bulk operations via API, Portal configuration.
Out of Scope: Engineering-side Jira/Linear configuration (see jira-expert/, linear-expert/). Strategic prioritization framework selection (see execution/prioritization-frameworks/). PRD authoring (see execution/create-prd/). Customer interview synthesis (see discovery/interview-synthesis/).
Limitations: API rate limits are workspace-wide; aggressive scripts can starve the UI. The Salesforce/HubSpot integration is one-way (CRM → Productboard) on a schedule, not real time. Drivers max out at 10 per workspace on most plans. Roadmap views are good for time-bound delivery (Releases) but weaker for outcome-based roadmaps (Now/Next/Later) — augment with execution/outcome-roadmap/. Per-Component access control is an Enterprise feature.
| Integration | Direction | What flows | |---|---|---| | jira-expert/ | Productboard ↔ Jira | Features push as Epics; status flows back; Sprint plan derived from prioritization | | linear-expert/ | Productboard ↔ Linear | Features push as Projects/Issues; Cycle assignment from Release dates | | execution/prioritization-frameworks/ | Bidirectional | Drivers operationalize RICE/ICE/WSJF; external scores import as custom fields | | execution/create-prd/ | Productboard → PRD | High-priority Features become PRDs; PRD link returns as custom field | | execution/outcome-roadmap/ | Bidirectional | Outcome themes map to Objectives; Releases give the delivery side | | execution/roadmap-communication/ | Productboard → Comms | Customer-facing Roadmap variants pull from Portal | | execution/customer-feedback-triage/ | Bidirectional | Triaged feedback clusters become Insights; inbox is the triage entry point | | execution/release-notes/ | Productboard → Release Notes | Completed Features in a Release become release-note line items | | senior-pm/ | Productboard → Portfolio | Driver scores and Feature distribution feed portfolio health | | notion-pm/ | Productboard ↔ Notion | Roadmap embeds into Notion; deep PRDs live in Notion linked from Features | | discovery/interview-synthesis/ | Bidirectional | Interview insights become Notes; high-evidence Features motivate interviews | | business-growth/customer-success/ | CS → Productboard | CS is a primary Insight source via Salesforce / Intercom integration |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-07 | fail→fail | 15,501 | 13,910 | -10% | 1 | 1 | 0% | 2,403 | 3,664 | +52% | 0 | 0 | — |
case-01 | fail→fail | 23,827 | 24,668 | +4% | 1 | 1 | 0% | 3,948 | 5,210 | +32% | 0 | 0 | — |
case-02 | fail→pass | 14,408 | 6,473 | -55% | 1 | 1 | 0% | 2,270 | 2,390 | +5% | 0 | 0 | — |
case-03 | fail→fail | 5,392 | 10,086 | +87% | 1 | 1 | 0% | 840 | 2,875 | +242% | 0 | 0 | — |
case-04 | fail→pass | 15,377 | 12,950 | -16% | 1 | 1 | 0% | 2,404 | 3,526 | +47% | 0 | 0 | — |
case-05 | fail→pass | 16,372 | 14,686 | -10% | 1 | 1 | 0% | 2,500 | 4,010 | +60% | 0 | 0 | — |
case-06 | pass→pass | 11,823 | 7,778 | -34% | 1 | 1 | 0% | 1,744 | 2,495 | +43% | 0 | 0 | — |
case-08 | fail→fail | 15,648 | 12,863 | -18% | 1 | 1 | 0% | 2,553 | 3,694 | +45% | 0 | 0 | — |
case-09 | fail→pass | 12,262 | 10,906 | -11% | 1 | 1 | 0% | 1,839 | 3,256 | +77% | 0 | 0 | — |
case-10 | pass→pass | 7,621 | 3,441 | -55% | 1 | 1 | 0% | 1,183 | 2,032 | +72% | 0 | 0 | — |
case-11 | pass→pass | 12,763 | 11,273 | -12% | 1 | 1 | 0% | 1,985 | 3,210 | +62% | 0 | 0 | — |
case-12 | fail→pass | 5,578 | 8,437 | +51% | 1 | 1 | 0% | 892 | 2,713 | +204% | 0 | 0 | — |
case-13 | pass→pass | 13,462 | 11,590 | -14% | 1 | 1 | 0% | 1,903 | 3,058 | +61% | 0 | 0 | — |
case-14 | fail→pass | 5,101 | 5,187 | +2% | 1 | 1 | 0% | 769 | 2,237 | +191% | 0 | 0 | — |
case-15 | pass→pass | 12,640 | 9,494 | -25% | 1 | 1 | 0% | 1,816 | 2,851 | +57% | 0 | 0 | — |
case-16 | pass→pass | 15,142 | 13,666 | -10% | 1 | 1 | 0% | 2,387 | 3,627 | +52% | 0 | 0 | — |
case-17 | pass→pass | 8,458 | 7,443 | -12% | 1 | 1 | 0% | 1,216 | 2,580 | +112% | 0 | 0 | — |
case-18 | pass→pass | 7,579 | 5,922 | -22% | 1 | 1 | 0% | 1,147 | 2,373 | +107% | 0 | 0 | — |
case-19 | pass→pass | 11,122 | 12,647 | +14% | 1 | 1 | 0% | 1,851 | 3,378 | +82% | 0 | 0 | — |
case-20 | pass→pass | 9,976 | 8,790 | -12% | 1 | 1 | 0% | 1,518 | 2,841 | +87% | 0 | 0 | — |
case-21 | pass→pass | 12,622 | 9,162 | -27% | 1 | 1 | 0% | 1,923 | 2,831 | +47% | 0 | 0 | — |
case-22 | pass→pass | 12,097 | 10,312 | -15% | 1 | 1 | 0% | 1,915 | 3,293 | +72% | 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. 22 cases were attempted. The headline lift of +27 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
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.