Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Guide the user to add a data source, connection, or API connector to a Canvas App via Power Apps Studio, then verify and continue. USE WHEN the user asks to add a data source, add a connection, add an API, add a connector, connect to SharePoint / Dataverse / SQL / Excel / OneDrive / Teams / Office 365, or any similar request to make new data available to the app. DO NOT USE WHEN the user is asking to list or describe existing data sources — call list_data_sources or list_apis directly instead.
.claude/skills/hashgraph-online-add-data-source/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-13 | ✗→✓ | ▲ Improved | 23% | 0% |
| case-21 | ✓→✗ | ▼ Worse | -21% | 0% |
| case-15 | ✓→✓ | ＝ Same ✓ | 59% | 0% |
| case-20 | ✓→✓ | ＝ Same ✓ | 386% | 0% |
| case-22 | ✓→✓ | ＝ Same ✓ | 86% | 0% |
Data sources, connections, and API connectors cannot be added by the coding agent — they must be added through the Power Apps Studio interface. This skill informs the user, guides them to add the connection in their Studio session, verifies it is available via the MCP server, and then continues with any pending work.
From the user's request, determine what they want to add and how to verify it:
| Request type | Verify with | |---|---| | Data source (SharePoint list, Dataverse table, SQL table, Excel file, OneDrive, etc.) | list_data_sources | | API / connector (Office 365 Users, Teams, custom connector, etc.) | list_apis | | Unclear | Both list_data_sources and list_apis |
Note what the user is trying to add so you can look for it by name in Phase 3.
Explain that this step requires action in their Studio session. Tell the user:
Use AskUserQuestion to pause until the user has completed the steps in Studio:
> "Please add the data source or connection in your Power Apps Studio session. Reply here when it's ready and I'll verify the connection before continuing."
Do not proceed until the user confirms they have added it.
Once the user confirms, call the appropriate MCP tools to check that the connection is now available:
list_data_sourceslist_apisScan the results for the connection the user added:
Found — Confirm to the user that the data source or connection is now visible to the agent. Include the exact name as returned by the MCP tool (this is how it must be referenced in Power Fx formulas). If it is a data source, call get_data_source_schema to retrieve its column names and types, and include the schema summary in your response so it is ready for use in edits. Similarly, for APIs/connectors, call describe_api to retrieve the available endpoints and their parameters, and include the schema summary in your response.
Not found — Inform the user that the connection is not yet visible. Ask them to:
Use AskUserQuestion to wait for their follow-up, then re-run Phase 3.
After the connection is confirmed:
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-15 | pass→pass | 12,406 | 6,656 | -46% | 1 | 1 | 0% | 1,010 | 1,605 | +59% | 0 | 0 | — |
case-01 | fail→fail | 3,953 | 7,921 | +100% | 1 | 1 | 0% | 679 | 1,330 | +96% | 0 | 0 | — |
case-02 | fail→fail | 14,796 | 8,229 | -44% | 1 | 1 | 0% | 1,721 | 1,320 | -23% | 0 | 0 | — |
case-03 | fail→fail | 7,637 | 4,092 | -46% | 1 | 1 | 0% | 976 | 1,366 | +40% | 0 | 0 | — |
case-04 | fail→fail | 9,779 | 17,456 | +79% | 1 | 1 | 0% | 726 | 1,303 | +79% | 0 | 0 | — |
case-05 | fail→fail | 2,704 | 17,083 | +532% | 1 | 1 | 0% | 354 | 1,775 | +401% | 0 | 0 | — |
case-06 | fail→fail | 16,996 | 8,904 | -48% | 1 | 1 | 0% | 1,845 | 1,470 | -20% | 0 | 0 | — |
case-07 | fail→fail | 14,927 | 8,535 | -43% | 1 | 1 | 0% | 1,686 | 1,418 | -16% | 0 | 0 | — |
case-08 | fail→fail | 14,318 | 8,094 | -43% | 1 | 1 | 0% | 1,340 | 1,342 | +0% | 0 | 0 | — |
case-09 | fail→fail | 11,414 | 9,619 | -16% | 1 | 1 | 0% | 1,711 | 1,183 | -31% | 0 | 0 | — |
case-10 | fail→fail | 11,131 | 11,377 | +2% | 1 | 1 | 0% | 1,074 | 1,197 | +11% | 0 | 0 | — |
case-11 | fail→fail | 14,719 | 8,336 | -43% | 1 | 1 | 0% | 1,691 | 1,269 | -25% | 0 | 0 | — |
case-12 | fail→fail | 12,971 | 9,387 | -28% | 1 | 1 | 0% | 1,101 | 1,558 | +42% | 0 | 0 | — |
case-13 | fail→pass | 10,943 | 11,366 | +4% | 1 | 1 | 0% | 1,500 | 1,844 | +23% | 0 | 0 | — |
case-14 | fail→fail | 13,690 | 14,713 | +7% | 1 | 1 | 0% | 1,646 | 1,111 | -33% | 0 | 0 | — |
case-16 | fail→fail | 13,749 | 12,855 | -7% | 1 | 1 | 0% | 1,306 | 1,718 | +32% | 0 | 0 | — |
case-17 | fail→fail | 10,802 | 3,793 | -65% | 1 | 1 | 0% | 2,018 | 1,133 | -44% | 0 | 0 | — |
case-18 | fail→fail | 15,901 | 10,121 | -36% | 1 | 1 | 0% | 1,831 | 1,174 | -36% | 0 | 0 | — |
case-19 | fail→fail | 15,583 | 9,825 | -37% | 1 | 1 | 0% | 1,858 | 1,164 | -37% | 0 | 0 | — |
case-20 | pass→pass | 9,226 | 13,678 | +48% | 1 | 1 | 0% | 705 | 3,428 | +386% | 0 | 0 | — |
case-21 | pass→fail | 14,530 | 16,954 | +17% | 1 | 1 | 0% | 1,682 | 1,332 | -21% | 0 | 0 | — |
case-22 | pass→pass | 9,332 | 12,332 | +32% | 1 | 1 | 0% | 1,609 | 2,999 | +86% | 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, and 12 counted toward the lift figure. The other 10 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 0 percentage points is the difference between those two pass rates over the 12 comparable cases. 3 cases got worse with the skill loaded, and they are 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.