Plugin check: Run node "${PLUGIN_ROOT}/scripts/check-version.js" - if it outputs a message, show it to the user before proceeding.
📋 Shared instructions: shared-instructions.md — read first.
References:
Add SharePoint
App root: before any project read or command, execute
app-working-directory.md.
Use its resolved absolute working_dir for every shell call and file tool,
including referenced commands and delegated skills; never inherit a prior cd.
Invocation scope: follow Data-source invocation scope
before the workflow below. Standalone calls approve only their SharePoint
schema/binding delta and verify it here; they do not start a full-app workflow.
Keep SharePoint list/library schemas in Connectors, not the Dataverse
Data Model. Forward supplied site/list/connection choices and ask only for missing
values; return scope changes to the current owner, or ask the standalone user,
before mutation. Reuse supplied approval only for the exact current data delta.
Removal branch: after resolving invocation scope, --remove or an approved list/library
binding removal executes
data-source-removal.md and
returns. Do not create a connection/list or run Steps 1-12 for removal; the
SharePoint list/library and its contents remain on the server.
Refresh branch: after resolving invocation scope, --refresh or an approved retained-list
service refresh executes Refresh a retained source
and returns before Steps 1-12. Preserve the exact --data-source-name and approved
binding identity; do not create a list/connection or run add-data-source.
Two paths: existing lists (skip to Step 6) or new lists (full workflow).
Workflow
- Check Memory Bank → 2. Plan → 3. Setup Graph API Auth → 4. Review Existing Lists → 5. Create Lists → 6. Get Connection ID → 7. Discover Sites → 8. Discover Tables → 9. Add Connector → 10. Configure → 11. Type-check → 12. Update Memory Bank
Step 1: Check Memory Bank
Check for memory-bank.md per shared-instructions.md.
Also confirm this is a mobile app:
cd -- '<working_dir>' || { echo "BLOCKED: cannot enter working_dir" >&2; exit 1; }
if [ ! -f power.config.json ] || [ ! -f app.config.js ]; then
echo "BLOCKED: working_dir is not an initialized app" >&2
exit 1
fi
Step 2: Plan
Telemetry checkpoint: plan_sharepoint_data_source
Reuse the supplied site/list and existing-vs-new decisions. Ask only for missing
values:
- Which SharePoint list(s) do they need?
- Do the lists already exist on their site, or do they need to create new ones?
For --plan-only or a planning-phase handoff, present the proposed site/list
bindings, any schema changes, and unresolved choices, then STOP before
implementation. Do not create connections/lists, generate services, or update
the app plan or memory-bank. Missing discovery access is not permission to
create a connection.
If lists already exist: Skip to Step 6.
If creating new lists:
- Ask about the data they need and design an appropriate schema
- Reuse existing lists when possible (don't duplicate)
- Enter plan mode with
EnterPlanMode, present the list designs with columns and types
- Get approval with
ExitPlanMode, unless the exact list/schema delta already
has current owner approval. Read-only discovery that changes that delta must
return for approval before Step 5.
Step 3: Setup Graph API Auth (if creating lists)
See api-authentication-reference.md for full details.
Set-Location -LiteralPath '<working_dir>' -ErrorAction Stop
az account show # Verify Azure CLI logged in
$api = Initialize-SharePointGraphApi -SiteUrl "https://<tenant>.sharepoint.com/sites/<site-name>"
$headers = $api.Headers
$siteId = $api.SiteId
Requires Sites.Manage.All permission.
Step 4: Review Existing Lists (if creating lists)
Always query existing lists first before creating:
Set-Location -LiteralPath '<working_dir>' -ErrorAction Stop
$existingLists = Invoke-RestMethod -Uri "https://graph.microsoft.com/v1.0/sites/$siteId/lists?`$select=id,displayName,description,list&`$filter=list/hidden eq false" -Headers $headers
See list-management-reference.md for Find-SimilarLists, Compare-ListSchemas, and Get-ListSchema functions.
Present findings to user with AskUserQuestion:
- Lists that can be reused (already exist with matching columns)
- Lists that need extension (exist but missing columns)
- Lists that must be created (no match found)
Step 5: Create Lists (if creating lists)
Telemetry checkpoint: create_sharepoint_lists
Print before starting:
"→ Creating SharePoint lists via Graph API (sequential per list, columns added after list exists)…"
Require explicit current approval for the exact create/extend operations before
writing; reuse matching owner approval rather than prompting twice. Use safe
functions from list-management-reference.md:
New-SharePointListIfNotExists
Add-SharePointColumnIfNotExists
Add-SharePointLookupColumn (for cross-list references)
Step 6: Get Connection ID
Apply connector-reference.md
before any lookup or creation. Use shared_sharepointonline as apiId.
- Supplied
--connection-id or approved connectionId: reuse the exact ID;
skip create-connection and /list-connections.
- Supplied
--connection-ref or approved connectionRef: retain the exact
reference for Step 9; skip create-connection and /list-connections.
- Missing binding only: use the approved solution-reference lookup when
solution-aware, or obtain explicit approval to create a connection with the
command below. Conflicting,
blank, or wrong-environment supplied bindings must be clarified, not replaced.
cd -- '<working_dir>' || { echo "BLOCKED: cannot enter working_dir" >&2; exit 1; }
$PA connection create --connector shared_sharepointonline --json
Only on the creation path, capture the returned connectionId. The following
picker examples use the ID path. With a reference, skip pickers when site/list
values are supplied; otherwise obtain the backing ID for that same reference or
ask for the missing choices. Do not create another connection for discovery.
If pa connection create cannot complete because browser-based connection creation is disabled or the connector needs interactive auth, direct the user to create one:
Open https://make.powerapps.com/environments/<environment-id>/connections → + New connection → search "SharePoint" → Create. Then provide the connection ID or rerun /list-connections shared_sharepointonline.
Step 7: Discover Sites
Skip this picker when the approved site URL is already supplied.
Print before starting:
"→ Discovering SharePoint sites accessible to this connection…"
cd -- '<working_dir>' || { echo "BLOCKED: cannot enter working_dir" >&2; exit 1; }
$PA connection list-datasets --connector shared_sharepointonline --connection-id '<connectionId>' --json
Present the discovered sites only when a site choice is missing.
If pa connection list-datasets fails or returns no results:
- Auth, wrong user, or multiple accounts: follow shared-instructions command-failure handling and retry once.
- Empty list: Confirm the connection ID is for a SharePoint Online connection and the user has access to at least one site. STOP if the list is empty after confirming.
Step 8: Discover Tables
Skip this picker for supplied list/library identities; reuse the approved values.
Print before starting:
"→ Discovering lists/document libraries on each selected site…"
For each selected site:
cd -- '<working_dir>' || { echo "BLOCKED: cannot enter working_dir" >&2; exit 1; }
$PA connection list-tables --connector shared_sharepointonline --connection-id '<connectionId>' --dataset '<site-url>' --json
Present the tables to the user and ask which ones they want to add. Suggest tables that look relevant to their use case. If lists were created in Step 5, they should appear here.
Step 9: Add Connector
Telemetry checkpoint: generate_sharepoint_data_source
Print before starting:
"→ Running pa app add data-source per list (sequential, ~10–20 seconds each)."
SharePoint is tabular: use --dataset and --table with the exact
binding selected in Step 6. Confirm approval covers this environment, connection
ID/reference, site, and list/library before registration. Ask the standalone
user to approve any newly resolved binding; return a changed child scope to its
owner. A picker selection alone is not execution approval.
Run only the applicable command:
cd -- '<working_dir>' || { echo "BLOCKED: cannot enter working_dir" >&2; exit 1; }
# Supplied/resolved connection ID
$PA app add data-source --connector shared_sharepointonline --connection-id '<connectionId>' --dataset '<site-url>' --table '<table-name>'
# Supplied/resolved connection reference
$PA app add data-source --connector shared_sharepointonline --connection-ref '<connectionRef>' --dataset '<site-url>' --table '<table-name>'
Run once per list or document library.
Inspect and return service signatures and usage guidance to the user or current
owner. Do not independently edit screens or launch another app workflow;
explicitly report that this data-source operation did not wire screens.
Read sharepoint-reference.md before writing any SharePoint code — column encoding, choice fields, and lookups have critical gotchas.
Use Grep to find methods in <working_dir>/src/generated/services/SharePointOnlineService.ts (generated files can be very large — see connector-reference.md).
Sample usage:
import { SharePointOnlineService } from '../../src/generated/services/SharePointOnlineService';
// Read items
const result = await SharePointOnlineService.GetItems({
dataset: 'https://contoso.sharepoint.com/sites/projects',
table: 'Project Milestones',
});
const items = result.value ?? [];
// Create item
await SharePointOnlineService.PostItem({
dataset: 'https://contoso.sharepoint.com/sites/projects',
table: 'Project Milestones',
item: { Title: 'Launch review', Status: 'Not Started' },
});
// Update item (SharePoint IDs are integers, not GUIDs)
await SharePointOnlineService.PatchItem({
dataset: 'https://contoso.sharepoint.com/sites/projects',
table: 'Project Milestones',
id: 42,
item: { Status: 'Done' },
});
Step 11: Type-check
Telemetry checkpoint: validate_sharepoint_integration
Print before starting:
"→ Regenerating connector schemas + running tsc to verify SharePoint services compile (~15–30 seconds)."
Native diff: npx --no-install tsc --noEmit instead of npm run build. Do NOT run platform-specific native build commands here.
pa app add data-source wrote new files into .power/schemas/sharepointonline/. Regenerate connectorSchemas.ts before type-checking so the new list is wired into the runtime schema map:
cd -- '<working_dir>' || { echo "BLOCKED: cannot enter working_dir" >&2; exit 1; }
npm run generate-schemas
npx --no-install tsc --noEmit
Fix TypeScript errors before proceeding.
Step 12: Update Memory Bank
Update memory-bank.md with: connector added, site URL, lists/libraries connected (or created), type-check status.
Note: No manual executor wiring needed. PowerAppsProvider in app/_layout.tsx handles SharePoint connector routing, connection resolution, and OAuth consent automatically at runtime.