edit-offline-profile

Use when the user wants to change ONE aspect of an existing offline profile (row scope for a table, column list, sync frequency) without re-running the full…

npx skills add https://github.com/microsoft/power-platform-skills --skill edit-offline-profile

Shared instructions: shared-instructions.md — read first.

References:

Edit Offline Profile

Re-run a single piece of an existing profile — change one table's row scope, update the column list, adjust sync frequency, or rename the profile. Avoids the cognitive cost of walking the full /setup-offline-profile wizard for a one-line change.

Scope: existing profile (read from <working_dir>/offline-profile.json or --profile-id); edits at the table-item granularity. To ADD a new table return to the owner for /add-table-to-offline-profile; deleting an entire profile is outside this helper's scope, not an automatic recovery.

Invocation context and root binding

This internal helper is always a child invocation. Before any project read or command, execute app-working-directory.md. Require the owner's absolute working_dir; --working-dir, when supplied, must identify that same native-canonical root. Missing, relative, or conflicting root context returns NEEDS_CONTEXT before project access; an inaccessible directory is BLOCKED. Never use the launch cwd or introduce another root resolver.

Receive MOBILE_APP_ORCHESTRATING=1, orchestrator, working_dir, phase, approved_scope, and supplied answers from the owner. Preserve the selected environment ID/URL/tenant, profile ID, table allowlist, item IDs, and exact scope/column/sync/name/description changes. Missing or conflicting scope returns NEEDS_CONTEXT, not another environment/profile/table selection. The marker is not approval: schema approval alone does not authorize profile edits, publication, deletion, or membership edits. Retain Step 3's single confirm unless supplied answers already approve that same current offline change. For --plan-only or a planning-phase handoff, return the current/proposed diff after read-only discovery and STOP before Apply, PATCH, publication, or artifact writes. A gate response cannot lift proposal-only mode.

Apply the canonical contract to every shell call and file tool, including version/auth checks, manifest/screen reads, commands copied from references, verification, recovery, and retries. Re-supply the absolute PLUGIN_ROOT and required context on each call; no prior cd, export, or shell variable persists. Use the canonical Bash guard below or its PowerShell equivalent. Replace placeholders with validated, shell-quoted invocation values; missing values return NEEDS_CONTEXT, never execute an unresolved placeholder. File tools use absolute <working_dir>/... paths; plugin reads/scripts use ${PLUGIN_ROOT}/.... Never search another app when a file is absent.

Workflow

  1. Verify project + locate profile → 2. Resolve target (which table / what to change) → 3. Show current vs proposed → Single confirm → 4. PATCH → 5. Publish → 6. Update artifacts → 7. Summary

Step 1 — Verify project + locate profile

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

Read <working_dir>/power.config.json and require a nonempty environmentId matching the owner's selected environment. Reuse matching complete environment context, or perform this non-persisting lookup from the bound root:

cd -- '<working_dir>' || { echo "BLOCKED: cannot enter working_dir" >&2; exit 1; }
environment_id=$(node -p "require('./power.config.json').environmentId || ''") || {
  echo "BLOCKED: unreadable power.config.json" >&2; exit 1;
}
if [ -z "$environment_id" ] || [ "$environment_id" != "<selected-environment-id>" ]; then
  echo "NEEDS_CONTEXT: selected environment does not match working_dir" >&2
  exit 1
fi
node "${PLUGIN_ROOT}/scripts/resolve-environment.js" "$environment_id" --no-cache --require-tenant

Require the returned environment ID, HTTPS URL, and tenant to match the owner; incomplete or conflicting context returns NEEDS_CONTEXT. Never drop --no-cache --require-tenant, change configuration, or switch auth/environment to recover a failure. Use that exact <envUrl> and <tenantId> for every request.

Profile ID resolution (same priority as /assign-offline-profile Step 1):

SourceUsed when
$ARGUMENTS --profile-id <guid>Explicit selection matching the owner's profile context
<working_dir>/offline-profile.json top-level profileIdDefault for /setup-offline-profile-created projects, matching the owner's profile context
OtherwiseReturn NEEDS_CONTEXT to the owner to select the intended profile; never pick another profile silently

STOP if no profile found: "Run /setup-offline-profile first."

Verify the selected live profile belongs to the selected environment. A conflicting snapshot/argument returns NEEDS_CONTEXT; a 404 or permission failure is BLOCKED, not permission to discover a substitute. Resolve each <itemId> only within this profile and verify its table matches the exact requested logical name.

<working_dir>/power.config.json is intentionally NOT consulted for the profile ID. That file is owned by pa app init. The local profile ID lives in <working_dir>/offline-profile.json only.

Step 2 — Resolve target (what to edit)

Parse $ARGUMENTS:

Flag patternEffect
--rename <new-name>Update profile name
--describe <text>Update profile description
--table <logical-name> --scope <0|1|2>Change one table's recorddistributioncriteria. Combine with --me, --team, --bu to set sub-flags when scope=2.
--table <logical-name> --sync <minutes>Change one table's syncintervalinminutes (range 5–1440)
--table <logical-name> --columns add:col1,col2 remove:col3Add or remove logical names from selectedcolumns. Comma-separated, both add/remove optional.
--table <logical-name> --columns resetReplace selectedcolumns with union of always-include + manifest lookups + screen-grep'd (re-runs the architect's Step 6 union for this table)

If no flags → interactive picker. AskUserQuestion with up-to-4 most likely edits:

  • "Rename profile"
  • "Change a table's scope" → next message asks which table
  • "Change a table's sync frequency" → next message asks which table + value
  • "Edit a table's column list" → next message asks which table + add/remove

Restrict the picker and explicit flags to the owner's exact current scope. --columns reset reads the selected table from <working_dir>/.datamodel-manifest.json or <working_dir>/docs/plan-artifacts/.datamodel-manifest.json and greps only absolute <working_dir>/app/ and <working_dir>/src/ paths. Keep the chosen absolute manifest path for the whole invocation; no other-app fallback. Changing the table, profile, or operation requires returning NEEDS_CONTEXT to the owner, not expanding an existing approval.

Step 3 — Show current vs proposed (single gate)

GET the current item state from Dataverse for any tables being edited:

cd -- '<working_dir>' || { echo "BLOCKED: cannot enter working_dir" >&2; exit 1; }
node "${PLUGIN_ROOT}/scripts/dataverse-request.js" "<envUrl>" GET \
  "mobileofflineprofileitems(<itemId>)?\$select=name,recorddistributioncriteria,recordsownedbyme,recordsownedbymyteam,recordsownedbymybusinessunit,syncintervalinminutes,selectedcolumns" \
  --tenant-id "<tenantId>"

Render a current/proposed diff:

Profile  : <name> (<id>)
Editing  : <table-logical-name> item

                Current     →    Proposed
Scope         : Org+me     →    All records
Sync (min)    : 10         →    30
Columns       : 14         →    16  (add: chnl_notes, chnl_actual_visit_date)

AskUserQuestion: "Apply this change? [Apply / Cancel]"

If Apply → continue. If Cancel → STOP.

Step 4 — PATCH

Build the PATCH body with only the fields that changed:

cd -- '<working_dir>' || { echo "BLOCKED: cannot enter working_dir" >&2; exit 1; }
node "${PLUGIN_ROOT}/scripts/dataverse-request.js" "<envUrl>" PATCH \
  "mobileofflineprofileitems(<itemId>)" --tenant-id "<tenantId>" \
  --body '{
    "recorddistributioncriteria": <new>,
    "recordsownedbyme": <new-bool>,
    "syncintervalinminutes": <new>,
    "selectedcolumns": "{\"Columns\":[...]}"
  }'

For profile-level edits (name / description):

cd -- '<working_dir>' || { echo "BLOCKED: cannot enter working_dir" >&2; exit 1; }
node "${PLUGIN_ROOT}/scripts/dataverse-request.js" "<envUrl>" PATCH \
  "mobileofflineprofiles(<profileId>)" --tenant-id "<tenantId>" \
  --body '{"name": "...", "description": "..."}'

Check returned HTTP statuses as well as process exit. A 403/PrivilegeCheckFailed is BLOCKED; do not change identity or profile to bypass permissions. Preserve every field and table outside the approved delta.

Step 5 — Publish

Use the targeted PublishXml recipe from shared/references/dataverse-offline-api.md §9:

cd -- '<working_dir>' || { echo "BLOCKED: cannot enter working_dir" >&2; exit 1; }
node "${PLUGIN_ROOT}/scripts/dataverse-request.js" "<envUrl>" POST \
  "PublishXml" --tenant-id "<tenantId>" --body '{
    "ParameterXml": "<publish><mobileofflineprofiles><mobileofflineprofile><profileId></mobileofflineprofile></mobileofflineprofiles></publish>"
  }'

Publishes only this profile, not the entire org's customizations. Avoids the 429 rate-limit storms on shared envs that the legacy PublishAllXml triggered.

On 400 / 0x80071141 "circular relationship" — same handling as /setup-offline-profile Step 8 (parse cycle path, prompt user to drop one association, retry). Fallback to PublishAllXml only if PublishXml returns an unexpected error other than the cycle case and the owner/user approves that broader publication; a profile-scoped approval does not cover it.

All linked recovery commands, re-GET verification, and retries must bind the same root and pass the same environment/tenant/profile/table/item context. Never retry on a different profile or bypass a permission failure. Confirm publication before Step 6.

If a linked recipe requires another offline helper, return the proposed change to the owner for approval first, then read the helper at its absolute ${PLUGIN_ROOT}/skills/<helper>/SKILL.md path. Forward:

Arguments: --working-dir '<working_dir>' --profile-id "<profileId>" --table "<table>" <exact-approved-change-flags>
Context: MOBILE_APP_ORCHESTRATING=1, orchestrator, working_dir, phase, approved_scope,
         selected environment ID/URL/tenant, profile ID, table allowlist, item ID,
         supplied answers, --plan-only (if present)

The recipient must enforce the same canonical root contract. Missing context returns NEEDS_CONTEXT; do not fall back to direct invocation or rediscovery.

Step 6 — Update artifacts

Re-read the changed item(s) in the same profile/environment and rewrite the matching entry in <working_dir>/offline-profile.json. When the edit changed a table's columns (--columns add:/remove:/reset), also refresh that entry's schemaColumns to the table's current full column set from the bound absolute manifest. This re-baselines the schema-reconciliation marker for the next ${PLUGIN_ROOT}/scripts/offline-profile-delta.js run; leave schemaColumns untouched for scope/sync/rename-only edits. Any delta check uses --project-root '<working_dir>' in its own guarded call. See offline-profile-reconciliation.md.

Append a one-line entry to <working_dir>/memory-bank.md ## Offline profile block:

edits:
  - { at: 2026-05-19T..., field: 'chnl_storevisit.syncintervalinminutes', from: 5, to: 10 }

Step 7 — Summary

✓ Edited profile.

  Field : chnl_storevisit.syncintervalinminutes
  From  : 5 min
  To    : 10 min
  Published: 2026-05-19T...

offline-profile.json + memory-bank.md updated.

Status code (final line)

  • DONE — change applied and published
  • DONE_WITH_CONCERNS: <list> — change applied but with caveats (publish timeout-then-success, etc.)
  • NEEDS_CONTEXT: <missing> — couldn't resolve target table or profile
  • BLOCKED: <reason> — auth or PATCH failure

What this skill does NOT do

  • Add a NEW table to the profile → use /add-table-to-offline-profile
  • Add a new association (relationship inclusion) → blocked on v0.2 selectedrelationshipsschema work; use maker portal in the meantime
  • Delete the whole profile → manual DELETE /mobileofflineprofiles(<id>) (cascade-deletes items + associations); see dataverse-offline-api.md §11
  • Migrate the profile between environments → use CloneMobileOfflineProfile action (v0.5 work)

More skills from microsoft

oss-growth
microsoft
OSS growth hacker persona
agent-framework-azure-ai-py
microsoft
Build Azure AI Foundry agents using the Microsoft Agent Framework Python SDK (agent-framework-azure-ai). Use when creating persistent agents with AzureAIAgentsProvider, using hosted tools (code interpreter, file search, web search), integrating MCP servers, managing conversation threads, or implementing streaming responses. Covers function tools, structured outputs, and multi-tool agents.
development
airunway-aks-setup
microsoft
Set up AI Runway on AKS — from bare cluster to running model. Covers cluster verification, controller install, GPU assessment, provider setup, and first deployment. WHEN: "setup AI Runway", "onboard AKS cluster", "install AI Runway", "airunway setup", "deploy model to AKS", "GPU inference on AKS", "KAITO setup on AKS", "run LLM on AKS", "vLLM on AKS", "set up model serving on AKS", "AI Runway controller".
devops
appinsights-instrumentation
microsoft
Guidance for instrumenting webapps with Azure Application Insights. Provides telemetry patterns, SDK setup, and configuration references. WHEN: how to instrument app, App Insights SDK, telemetry patterns, what is App Insights, Application Insights guidance, instrumentation examples, APM best practices.
devops
applicationinsights-web-ts
microsoft
Instrument browser/web apps with the Application Insights JavaScript SDK (@microsoft/applicationinsights-web). Use for Real User Monitoring (RUM) — page views, clicks, AJAX/fetch dependencies, exceptions, custom events, and browser-side GenAI agent traces correlated to backend OpenTelemetry traces. Covers SDK Loader Script and npm setup, framework extensions (React, React Native, Angular), Click Analytics, telemetry initializers, and OTel GenAI semantic conventions for agent/tool/model spans emitted from the browser.
devops
azure-ai-anomalydetector-java
microsoft
Build anomaly detection applications with Azure AI Anomaly Detector SDK for Java. Use when implementing univariate/multivariate anomaly detection, time-series analysis, or AI-powered monitoring.
development
azure-ai-language-conversations-py
microsoft
Implement Conversational Language Understanding (CLU) using the azure-ai-language-conversations Python SDK. Use when working with ConversationAnalysisClient to analyze conversation intent and entities, building NLP features, or integrating language understanding into applications.
development
azure-ai-ml-py
microsoft
Azure Machine Learning SDK v2 for Python. Use for ML workspaces, jobs, models, datasets, compute, and pipelines. Triggers: "azure-ai-ml", "MLClient", "workspace", "model registry", "training jobs", "datasets".
development