Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use the OpenProject plugin MCP tools to inspect projects and manage work packages in a configured OpenProject instance.
.claude/skills/hashgraph-online-openproject/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | -57% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -52% | 0% |
| case-14 | ✗→✓ | ▲ Improved | -79% | 0% |
| case-15 | ✗→✓ | ▲ Improved | -40% | 0% |
| case-16 | ✗→✓ | ▲ Improved | -63% | 0% |
Use this skill when the user wants to read or update data in OpenProject.
openproject server.server_info or get_current_user.first with list_projects and list_project_types.
server_info.list_projects.list_project_types.search_work_packages or get_work_package to gather context beforemaking changes.
create_work_package, update_work_package, andcomment_on_work_package for mutations.
The plugin reads:
OPENPROJECT_BASE_URLOPENPROJECT_API_TOKENOPENPROJECT_ACCESS_TOKENOpenProject documents API v3 authentication with Bearer tokens, API tokens, and OAuth2. A personal API token in OPENPROJECT_API_TOKEN is the simplest setup.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-23 | pass→pass | 73,841 | 67,271 | -9% | 1 | 1 | 0% | 1,967 | 1,060 | -46% | 0 | 0 | — |
case-09 | pass→pass | 10,498 | 8,054 | -23% | 1 | 1 | 0% | 763 | 679 | -11% | 0 | 0 | — |
case-01 | fail→fail | 13,161 | 30,663 | +133% | 1 | 1 | 0% | 815 | 617 | -24% | 0 | 0 | — |
case-02 | fail→fail | 12,069 | 19,439 | +61% | 1 | 1 | 0% | 284 | 725 | +155% | 0 | 0 | — |
case-03 | fail→fail | 10,629 | 15,103 | +42% | 1 | 1 | 0% | 497 | 564 | +13% | 0 | 0 | — |
case-04 | fail→fail | 13,103 | 15,146 | +16% | 1 | 1 | 0% | 1,142 | 1,141 | -0% | 0 | 0 | — |
case-05 | fail→fail | 22,646 | 50,937 | +125% | 1 | 1 | 0% | 505 | 1,654 | +228% | 0 | 0 | — |
case-06 | fail→fail | 47,696 | 22,660 | -52% | 1 | 1 | 0% | 2,253 | 2,218 | -2% | 0 | 0 | — |
case-07 | fail→pass | 15,130 | 33,225 | +120% | 1 | 1 | 0% | 1,696 | 731 | -57% | 0 | 0 | — |
case-08 | pass→pass | 16,912 | 9,480 | -44% | 1 | 1 | 0% | 1,735 | 901 | -48% | 0 | 0 | — |
case-10 | pass→pass | 5,458 | 3,490 | -36% | 1 | 1 | 0% | 848 | 684 | -19% | 0 | 0 | — |
case-11 | fail→fail | 14,399 | 15,238 | +6% | 1 | 1 | 0% | 1,238 | 724 | -42% | 0 | 0 | — |
case-12 | pass→pass | 5,311 | 2,142 | -60% | 1 | 1 | 0% | 829 | 558 | -33% | 0 | 0 | — |
case-13 | fail→pass | 8,784 | 8,023 | -9% | 1 | 1 | 0% | 1,399 | 675 | -52% | 0 | 0 | — |
case-14 | fail→pass | 20,487 | 7,871 | -62% | 1 | 1 | 0% | 2,780 | 592 | -79% | 0 | 0 | — |
case-15 | fail→pass | 11,744 | 8,204 | -30% | 1 | 1 | 0% | 1,052 | 626 | -40% | 0 | 0 | — |
case-16 | fail→pass | 21,004 | 2,121 | -90% | 1 | 1 | 0% | 1,742 | 650 | -63% | 0 | 0 | — |
case-17 | pass→pass | 12,242 | 8,567 | -30% | 1 | 1 | 0% | 1,271 | 810 | -36% | 0 | 0 | — |
case-18 | pass→pass | 5,163 | 2,774 | -46% | 1 | 1 | 0% | 770 | 632 | -18% | 0 | 0 | — |
case-19 | pass→pass | 11,128 | 7,422 | -33% | 1 | 1 | 0% | 710 | 595 | -16% | 0 | 0 | — |
case-20 | pass→pass | 34,631 | 16,397 | -53% | 1 | 1 | 0% | 2,857 | 2,730 | -4% | 0 | 0 | — |
case-21 | pass→pass | 27,561 | 111,659 | +305% | 1 | 1 | 0% | 3,063 | 2,534 | -17% | 0 | 0 | — |
case-22 | pass→pass | 21,660 | 25,493 | +18% | 1 | 1 | 0% | 2,424 | 2,568 | +6% | 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, and 18 counted toward the lift figure. The other 5 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +22 percentage points is the difference between those two pass rates over the 18 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.