OverviewHistoryStatsSecurity
npx skills add ...
Documentation
SKILL.md
npx skills add microsoft/power-platform-skills --skill plan-alm
Creates an ALM (Application Lifecycle Management) plan for deploying a Power Pages site across environments. Gathers your promotion strategy, target environments, and approval requirements upfront, then generates a visual HTML plan document for your review and approval. **plan-alm does not deploy anything itself** — it is a planner. After you approve the plan, run the individual ALM skills (setup-solution, setup-pipeline, deploy-pipeline, or export-solution/import-solution); each detects the approved plan and executes the right step in order, keeping the plan updated as it runs. Use when asked to: "plan my alm", "set up alm", "create deployment plan", "plan my deployments", "help me deploy to multiple environments", "set up promotion strategy", "create cicd plan", "plan site promotion", "help me go to production", "set up pipeline for my site".
npx skills add microsoft/power-platform-skills --skill plan-alm
Plugin check: Run
node "${PLUGIN_ROOT}/scripts/check-version.js"— if it outputs a message, show it to the user before proceeding.
A 4-phase planner that gathers ALM strategy from the user, generates an HTML deployment plan, and gets approval. It does not execute anything — execution is delegated to the individual ALM skills, which the user runs afterward.
This skill detects the current project state (existing solution, pipeline), asks targeted questions about the desired promotion strategy (Power Platform Pipelines or Manual export/import), generates a visual docs/alm-plan.html, and gets user approval. The four phases are: Phase 1 — Detect, Phase 2 — Gather strategy, Phase 3 — Generate plan, Phase 4 — Approve & save.
plan-alm never deploys. The plan's steps[] array records the recommended execution sequence. After approval, the user invokes the individual skills — setup-solution, setup-pipeline (or export-solution), and deploy-pipeline (or import-solution) — in that order. Each of those skills detects the approved plan via its Phase 0 gate, proceeds without re-nagging, and refreshes the plan on completion. This separation is deliberate: it keeps plan-alm safe to run unattended (e.g. under autopilot) because no single answer can trigger an irreversible deployment.
Do NOT create tasks at the start — strategy is unknown until Phase 2 completes. Create both tasks in Phase 3 once the strategy is determined.
Do NOT create tasks yet. Use natural language progress reporting only during this phase.
Steps:
Detect prior ALM deferral for this project. Before any discovery work, check whether the project root contains a .alm-deferred marker file. The marker is written by users who explicitly opted ALM-skill validators out of "missing artifacts" warnings (e.g. "this site is handled separately" or "ni-dev — no ALM"). If a user is now invoking plan-alm, we should surface that the marker is present and ask what to do, rather than silently proceeding (which would build a plan the user previously decided not to maintain) or silently removing the marker (which would re-enable nags on every other ALM skill).
Use
--no-heartbeathere: this is a read-only deferral check by the planner, not an execution-skill Phase 0 gate. Without it,check-alm-plan.jswould promote an already-Approvedplan toIn Execution(and refresh the heartbeat) just because you re-openedplan-alm— but re-planning isn't execution. Execution skills call it without--no-heartbeatso the first one to run does theApproved → In Executionpromotion.
🚦 Gate (progress · plan-alm:1.deferral):
.alm-deferredmarker present — continue and remove, continue and keep marker, or cancel. Determines whether downstream ALM skills resume gate enforcement.
The helper returns { deferred, deferral, ... }. If deferred === true, read the deferral reason (deferral.reason or the raw marker text) and ask via AskUserQuestion:
"This project has an
.alm-deferredmarker —{reason}. ALM was previously deferred here, so the other ALM skills (setup-solution,setup-pipeline,deploy-pipeline, …) skip their plan-completeness checks for this project. How would you like to proceed?"
| Question | Header | Options |
|---|---|---|
| How would you like to proceed? | ALM deferral marker | Continue planning and remove the marker (Recommended), Continue planning but keep the marker (record deferral context in plan), Cancel |
.alm-deferred (the user is re-engaging with ALM). Set DEFERRAL_CLEARED = true and proceed to step 1.DEFERRAL_PRESERVED = true and DEFERRAL_REASON = {reason}. Proceed to step 1. Surface a one-line note in the Phase 1 step 9 user report (e.g. "Note: .alm-deferred is preserved — other ALM skills will continue to skip plan-completeness checks for this project.") so the user remembers the marker remains in effect after planning.If deferred === false, skip this step silently and proceed to step 0b.
0b. Offer to approve an existing Draft in place (skip re-planning). The same check-alm-plan.js output from step 0 also carries exists and planStatus. When exists === true and planStatus === "Draft", the user already has a saved Draft plan — offer to approve it directly instead of regenerating the whole plan. (This is the only Draft→Approved path; without it, approving a draft means a full re-plan.)
🚦 Gate (plan · plan-alm:1.approve-draft): An existing Draft plan was found — approve it in place (no re-plan), re-plan from scratch, or cancel. Approving here writes the status via
set-plan-status.jsand exits without re-running discovery; no deployment is triggered.
Ask via AskUserQuestion:
"This site already has an ALM plan saved as Draft (
docs/alm-plan.html). What would you like to do?"
| Question | Header | Options |
|---|---|---|
| What would you like to do? | Existing draft plan | Approve this draft now — no re-plan (Recommended), Re-plan from scratch, Cancel |
Approve this draft now (Recommended) → capture the approver using the Phase 4 approver-capture procedure (the always-interactive prompt with git/OS-name prefill), then write the status atomically with the helper:
Commit (git add docs/alm-plan.html docs/.alm-plan-data.json && git commit -m "Approve ALM plan for {siteName}"), run skill tracking (Phase 4 finalize), print the Phase 4 next-steps guidance, and exit. Do not continue to step 1 — there is nothing to re-plan.
Re-plan from scratch → proceed to step 1 (the rest of Phase 1 regenerates the plan; Phase 4 saves the new version).
Cancel → exit cleanly (leave the Draft as-is).
If exists === false, or planStatus is anything other than "Draft" (Approved / In Execution / Completed / null), skip this step silently and proceed to step 1.
Resolve the site identity from the local project. .powerpages-site/website.yml is the source of truth for websiteRecordId and siteName, and it is present for both Power Pages site types:
/power-pages:create-site and downloaded with pac pages download-code-site. These also have a powerpages.config.json and SPA source (src/, build output in dist//build/).pac pages download --modelVersion 1|2. These have no powerpages.config.json; instead .powerpages-site/ holds the config tree (web-pages/, web-templates/, content-snippets/, …) plus a .powerpages-site/.portalconfig/ manifest pair. There is no local build output.Resolution order (first match wins):
.powerpages-site/website.yml (preferred, present for every downloaded/deployed site) — read with the Read tool and extract:
id field → websiteRecordIdname field → siteName (the file uses short keys; it is name:, not adx_name:)powerpages.config.json (fallback — code/SPA sites only; used during plugin development from this repo root or for sites scaffolded but not yet deployed) — read siteName and websiteRecordId.Determine SITE_TYPE (recorded in planData as siteType and used to skip SPA-only assumptions below; it is a data field in docs/.alm-plan-data.json, not rendered in the HTML):
declarative when .powerpages-site/.portalconfig/ exists, or .powerpages-site/website.yml resolved while no powerpages.config.json is present. (This value was formerly data-model; plans written before the rename may still carry data-model, which is equivalent.)code when powerpages.config.json is present.If neither marker is found, stop with:
"No Power Pages site found in the current directory. Run this skill from your site project root — that's where
.powerpages-site/lives afterpac pages download-code-site(code/SPA site) orpac pages download --modelVersion 2(enhanced data-model site). If you haven't created the site yet, run/power-pages:create-sitefirst."
environmentUrl is always re-confirmed from pac env who in step 4 — it does not need to come from either source.
Check for .solution-manifest.json in the project root:
SOLUTION_DONE = true if found, false otherwisesolution.uniqueName and store as SOLUTION_UNIQUE_NAMECheck for docs/alm/last-pipeline.json in the project root:
PIPELINE_DONE = true if found, false otherwisepipelineName and stages[] for later useRun silently:
Capture the environment URL and display name. Store as DEV_ENV_URL and DEV_ENV_NAME. The URL label varies by PAC version: current PAC (2.8.x) prints it under Org URL:; older builds used Environment URL: — read whichever is present (there is no Environment URL: line on 2.8.x, so do not look only for that label). The display name is the Friendly Name: / Connected to... value. If you can't parse it reliably, leave DEV_ENV_URL empty — Step 6's verify-alm-prerequisites.js resolves the authoritative URL from pac env who via the shared getEnvironmentUrl() helper (which matches both labels) and returns it as .envUrl.
Run silently:
Store the JSON array as ENV_LIST for pre-filling environment URLs in Phase 2. (This helper parses pac env list; the old pac env list --output json is invalid on current PAC CLI — pac env list only accepts --filter — so the helper exists to produce the JSON the table form doesn't. It prints [] and exits 0 if PAC is unauthenticated, so pre-fill simply degrades to manual entry.) Each entry is { displayName, environmentId, environmentUrl, uniqueName, active }.
Acquire dev environment token (silently):
Store .token as DEV_TOKEN and .userId as userId.
Track plan quality. Initialize a PLAN_QUALITY accumulator to "complete" at the start of Phase 1. If this token acquisition fails (auth error), set DEV_TOKEN = null, set PLAN_QUALITY = "degraded", and record the cause (e.g. "dev-environment auth failed — contents/size/host discovery skipped") — then continue. Contents discovery is skipped gracefully, but the resulting plan is built on partial inputs; Phase 3 surfaces this as a prominent risk so the user reviews before executing. (There is no execute path to block here — plan-alm only plans — but a degraded plan must be visibly flagged.)
6b. Environment-match guard — confirm pac env who points at the project's environment before running discovery. DEV_ENV_URL comes from whatever environment PAC happens to be connected to, which is not guaranteed to be the project's. If it isn't, every query in Steps 7–12 runs against the wrong environment and silently produces a degraded plan (zero or wrong site settings, wrong size, wrong host) that looks valid. Cross-check both signals available:
Discover and classify site settings (if DEV_TOKEN is available and websiteRecordId is known):
Use Node.js https module to query. Paginate via @odata.nextLink — sites with > 500 settings would otherwise silently truncate, dropping tier classifications and underreporting plannedEnvVarCount. Send Prefer: odata.maxpagesize=5000 so Dataverse emits the continuation link, then loop until exhausted:
On each response, append value[] to the running array. If @odata.nextLink is present, GET that URL with the same headers (no need to re-add the filter — the nextLink already encodes the query). Stop when the response has no @odata.nextLink. Cap at 100 iterations for safety.
Classify the returned settings using ${PLUGIN_ROOT}/scripts/lib/classify-site-settings.js — the single source of truth for the credential regex and tier mapping shared with setup-solution Phase 5. Either pipe the JSON array of {name, value} rows into the script's stdin (CLI mode) or require() it inline:
Output (the four-bucket shape that downstream phases + setup-solution consume directly):
Tier semantics in plain English (so reviewers reading the plan know what each bucket implies):
credentialNeedsDecision) — credential-style names. Setup-solution Phase 5.4.C runs a single bulk prompt: auto-classify by name (Secret-typed env var for *Secret/*Password/*ApiKey/*AppKey; String-typed for *Id/*ConsumerKey), all-as-Secret, all-as-String, skip-all, or pick-per-credential.promoteToEnvVar) — auth config with a dev value. Setup-solution Phase 5.4.A asks which to back with env vars so each stage can use different values.authNoValue) — auth config with no dev value yet. Added to the solution as-is; user sets the value in each target env after deployment.keepAsIs) — everything else. Added unchanged.If the OData query fails or the helper errors out, set SITE_SETTINGS_DATA = null and continue — the plan still renders, it just can't break down site settings by tier.
Build SOLUTION_CONTENTS_DATA:
If SOLUTION_DONE = false and manifest is absent, tables and botComponents will be empty arrays — the plan will show a note that they will be discovered during setup-solution.
Report to user:
Estimate solution size and evaluate the split decision tree. First ensure the ALM artifacts directory exists (all .alm-* and last-* artifacts live under docs/alm/ to keep the project root uncluttered):
Run the estimate helper to classify the site across size, component count, schema heaviness, web file aggregate, and env var count. Use the tmp-file write pattern — if the estimator fails, a prior good docs/alm/alm-size-estimate.json is preserved instead of being overwritten with an empty/partial file. When SOLUTION_DONE = true (a .solution-manifest.json exists), pass --solutionId {solutionId} so the env var count is scoped to the target solution — without it, the estimator falls back to a publisher-prefix tenant-wide query and overcounts whenever the prefix is shared across projects (the common new_ / cr5fe_ regression):
When SOLUTION_DONE = false, omit --solutionId; the estimator's output will include envVarCountScope: "publisher-prefix" to signal the wider scope, and the renderer surfaces this caveat in the Env Variables tab so reviewers know the number reflects the tenant view, not a specific solution. --projectRoot "." enables the disk cross-check — the estimator walks the local build output (dist/, public-output/, build/, .output/) and surfaces webFilesDiskMeasuredMB. When that number is much larger than the Dataverse-measured webFilesAggregateMB, the estimator flips truncationSuspected: true with a warning — file-typed columns whose bytes aren't returned by $select=content are the usual cause and the plan should trust the disk number.
SITE_TYPE = "declarative"(EDM/standard data-model) sites have no build output, so the disk cross-check finds nodist//build/directory andwebFilesDiskMeasuredMBstaysnull— this is expected, not a problem. Web files for declarative sites live as records under.powerpages-site/web-files/and are measured via the Dataverse query, so the size estimate is still valid; there's simply no SPA bundle on disk to cross-check against. Pass--projectRoot "."regardless — it's a harmless no-op for these sites. Then run the decision tree (same tmp-file pattern):bash node "${PLUGIN_ROOT}/scripts/lib/compute-split-plan.js" \ --estimate ./docs/alm/alm-size-estimate.json \ --projectRoot "." \ --siteName "{siteName}" \ --publisherPrefix "{publisherPrefix}" > ./docs/alm/alm-split-plan.json.tmp \ && mv ./docs/alm/alm-split-plan.json.tmp ./docs/alm/alm-split-plan.jsonIf either command exits non-zero, stop and report the stderr message to the user. Do not proceed to Q1b in Phase 2 without a valid split plan. Store the output asSPLIT_PLAN. Fields to read:splitStrategy,proposedSolutions[],appliedStrategies[],assetAdvisory,sizeAnalysis,recommendations[].
10b. Enumerate environment variable definitions (runs whenever DEV_TOKEN is available — the size estimator gives a count but not per-variable metadata).
Pre-plan completeness check (only runs when SOLUTION_DONE = true).
Before the user approves a plan, verify the existing solution already covers everything on the live site. Components created after the last /power-pages:setup-solution run (server logic from add-server-logic, flows from add-cloud-flow, env vars from configure-env-variables or setup-auth) are silently excluded from any plan built on top of a stale solution.
Run the shared discovery helper against the source environment:
Parse stdout and evaluate missing.*:
All missing.* arrays empty → report "Solution contents match the site — proceeding with fresh plan inputs." Continue to Phase 2.
Any non-empty missing.* array → report a compact summary:
"Your solution is missing {N} component(s) that exist on the site:
- {X} site components (e.g. {first 3 names})
- {L} site languages (powerpagesitelanguage — required; without these the target site silently fails to render post-auth)
- {Y} cloud flows
- {Z} environment variable definitions
- {W} custom tables
A plan built now will ignore these components. How would you like to proceed?"
Always render the site languages line when missing.siteLanguages.length > 0, even when other categories are zero — this gap was a recurring silent-failure mode before discover-site-components started enumerating powerpagesitelanguages. See references/solution-api-patterns.md for the 3-entity model.
🚦 Gate (progress · plan-alm:1.completeness): Completeness check found gaps vs live site. Sync first, plan with gaps recorded, or cancel.
Ask via AskUserQuestion:
| Question | Header | Options |
|---|---|---|
Run /power-pages:setup-solution in sync mode to adopt the missing components before planning? | Completeness Check | Yes — sync first (Recommended), No — plan with current solution contents, Cancel |
/power-pages:setup-solution (auto-detects the existing manifest and enters sync mode). After it completes, re-run the discovery helper; if missing.* is now empty proceed to Phase 2, otherwise repeat the prompt.KNOWN_GAPS so Phase 3 can surface it in the plan HTML's Risks section, then continue.Why this exists: the same check runs at export (
export-solutionPhase 2.5) and deploy (deploy-pipelinePhase 3.5). Adding it here catches gaps at the earliest possible gate — before the user invests time reviewing a plan built on stale inputs. See AGENTS.md → ALM-aware by default.
Skip when
SOLUTION_DONE = false: if there is no manifest yet, there is nothing to be stale against — Phase 2 Q1 will handle first-time solution setup.
Run host resolution (PP Pipelines path only — runs after the completeness check).
Skip rule: if PIPELINE_DONE = true, skip this step entirely — the host info comes from docs/alm/last-pipeline.json. Only fresh-pipeline projects need resolution.
Acquire a BAP-audience access token (the BAP API uses a different audience than Dataverse):
Capture the output as BAP_TOKEN. If acquisition fails, set HOST_RESOLUTION = { status: 'DetectionFailed', error: '<stderr>' } and skip the detect call.
Run the detect-only wrapper. Use the same tmp-file-then-mv pattern as Phase 1 step 10 so a prior good docs/alm/alm-host-resolution.json is preserved if the script fails mid-write. Pass --skus Production,Sandbox,Trial so trial-license and developer tenants see their eligible envs in the env-first menu (the helper's default is Production,Sandbox; we widen to include Trial here because plan-alm's NoHost branch always offers an existing-env install path that Trial envs can take, even though Trial envs cannot use the create-new fast-path):
Note:
ensure-pipelines-host-detect.jsis a detection-only wrapper theensure-pipelines-hostskill exposes for orchestrators. It runs Phases 1.0 (cache fast-path) + 2 (resolution order including tenant-wide enumeration) + 5 (verify if a host is found) of that workflow, but never enters Phase 3 (decision tree) or Phase 4 (provisioning). Output matches thedocs/alm/last-host-check.jsonschemaVersion 2 withactionTaken: "none"always.
Failure handling: if the detection script exits non-zero, set HOST_RESOLUTION = { status: 'DetectionFailed', error: '<stderr>' } and continue. Phase 2 Q4 falls back to today's "enter URL manually" branch.
On success, parse docs/alm/alm-host-resolution.json and store as HOST_RESOLUTION (mapping the wrapper's field names into the plan-alm shape):
Report a single line:
or, when no URL is set yet:
Ask questions in sequence. Solution is always Q1 — it is the prerequisite for all other steps. Branch after Q2 based on promotion strategy selection.
Log every major decision. As each decision is made below (Q1 solution, Q1b split/override, Q2 strategy, Q3 stages/targets, Q4 host, Q5 approval mode, Q5 manual export type), append to a DECISIONS_LOG array: { field, value, source } where source = "default" when the recommended/auto value was accepted without an active change, or "explicit" when the user picked a non-default option. Phase 4 renders a "Decisions defaulted (please review)" section from this log so a reviewer can see at a glance which choices were defaults vs. deliberate picks (closes the "I never agreed to managed export" gap). This adds no new prompts — it only records what the existing prompts produced.
If SOLUTION_DONE = true (manifest found in Phase 1):
🚦 Gate (plan · plan-alm:2.q1-existing): Existing solution found — reuse it (skip setup-solution) or create new (run setup-solution).
Ask via AskUserQuestion:
"A Dataverse solution is already configured for this site: {SOLUTION_UNIQUE_NAME}. Use this existing solution?"
Options:
setup-solution will be skipped in the planSOLUTION_DONE = false; setup-solution will runIf SOLUTION_DONE = false (no manifest found):
Tell the user (not via AskUserQuestion — informational only):
"No Dataverse solution is set up for this site yet.
setup-solutionwill be the first step in your plan. The publisher prefix you choose during setup is irreversible — choose carefully."
🚦 Gate (plan · plan-alm:2.q1-fresh): No existing solution — include setup-solution in plan, or accept a user-supplied unique name.
Ask via AskUserQuestion:
"Ready to include solution setup in the plan?"
Options:
SOLUTION_DONE = true, SOLUTION_UNIQUE_NAME = user inputRECOMMEND_SPLIT = true)🚦 Gate (plan · plan-alm:2.q1b-split): Follow recommended split strategy, override to single, accept Asset Advisory first, or show migration guidance.
The decision tree from Phase 1 Step 10 recommended splitting into multiple solutions. Ask via AskUserQuestion:
"Based on the site size and component analysis, the recommended approach is {splitStrategy} — {N} solutions instead of one. Do you want to follow this recommendation?"
Options:
proposedSolutions[] from the decision tree. setup-solution will create all N solutions.setup-solution creates one solution with all components.assetAdvisory.candidates.length > 0) — user commits to externalizing the flagged assets. Recompute size excluding those files, re-run the decision tree, present the new recommendation..solution-manifest.json is found and does not match the recommendation) — produce docs/alm-migration-plan.md and exit. Do not execute.If option 1: continue with proposedSolutions.
If option 2 — Keep as a single solution anyway: this overrides a data-driven recommendation that's frequently right. Before honoring the override, re-surface the tier signals so the user is making an informed choice, not a one-click dismissal. Read from SPLIT_PLAN.sizeAnalysis:
🚦 Gate (consent · plan-alm:2.q1b-override): Override the data-driven split recommendation to keep as single solution. Free-text
overrideReasonfollows on Yes.
Then ask via AskUserQuestion:
| Question | Header | Options |
|---|---|---|
| Still want to keep as a single solution? | Override confirmation | No — use the recommended {splitStrategy} split (Recommended), Yes — override anyway and note the reason, Cancel — re-think the strategy |
overrideReason via a follow-up AskUserQuestion ("Briefly: why is single-solution the right call for this site?"). Record overrideReason and overrideConfirmedSignals (the tier-classified signals shown above) in the plan. Only then override SPLIT_PLAN.proposedSolutions to the single-solution structure for rendering.Why the friction: in field-reported sessions, "keep as single anyway" was a one-click override and turned out to be the single most common path to a wrong recommendation. The re-confirmation isn't there to talk the user out of it — it's there to make sure the override is informed and the reason gets recorded for audit. Override-with-recorded-reason is fully respected; the gate only blocks the silent click-through.
If option 3: subtract advisory candidate sizes from the estimate, re-run compute-split-plan.js, re-present.
If option 4: write docs/alm-migration-plan.md (see the spec doc solution-splitting-logic.md §7), commit it, mark plan as Deferred, exit.
🚦 Gate (plan · plan-alm:2.q2-strategy): Pick promotion strategy — PP Pipelines (Recommended), manual export/import, existing pipeline, or help-me-decide. Branches the rest of the plan.
Recommendation prelude (shown inline in the prompt body). Before listing the options, surface the recommendation and any state that affects it, so the user sees it up front instead of having to drill into option 4:
Recommendation: Power Platform Pipelines for ongoing CI/CD — automated promotion, approval gates, and deployment history in one place. Manual export/import is intended for one-off migrations only; repeating it for every change loses the audit trail and silently allows version-skew bugs.
(Conditional, append when
HOST_RESOLUTION.resolutionStatus∈ {AvailableUsingCustomHost,AvailableUsingCustomHostByAdminDefault,AvailableUsingPlatformHost,AvailableUnboundCustomHost}): A pipelines host already exists in your tenant at{HOST_RESOLUTION.finalHostEnvUrl}(Pipelines v{HOST_RESOLUTION.pipelinesSolutionVersion}). PP Pipelines requires no new infrastructure for this project.
Ask via AskUserQuestion:
"How do you want to promote your solution between environments?"
Options:
Record the pick in DECISIONS_LOG ({ field: "strategy", value, source }).
If option 4 selected: Re-print the recommendation prelude above, then re-ask Q2 with only options 1–3.
If option 3 selected: Read docs/alm/last-pipeline.json, confirm pipeline name and stages, then skip to Phase 3 (generate plan) with strategy = pp-pipelines, PIPELINE_DONE = true.
🚦 Gate (plan · plan-alm:2.q3-stages): Pick how many deployment stages — Staging only / +Production / Production directly / Custom.
Q3: Ask via AskUserQuestion:
"How many deployment stages do you want in this pipeline?"
Options:
If option 4: accept free-text description (via "Other") and build a stage list from the response.
Store stages as PP_STAGES (array of { label, envUrl, envName, type }). Dev is always the source.
For each stage, populate envName from ENV_LIST (gathered in Phase 1 Step 5 via list-environments.js). Match by URL origin (lowercase, trailing slash stripped, path/query ignored) against each entry's environmentUrl and copy the entry's displayName into envName. When no match is found — usually because the user pasted a custom URL via "Other" — leave envName unset; the renderer falls back to showing the URL alone in the stage card. The renderer puts envName between the stage label and the URL (e.g. Staging / Supplier Portal Staging / https://orgd6a9894f.crm5.dynamics.com/) so reviewers recognize the env at a glance and the URL stays available as a one-click jump-to-env. Set type: "source" for the dev/source stage and type: "target" for every downstream stage so the renderer applies the active-stage styling correctly.
Reconcile the chosen stages against an existing pipeline (soft warning, PP path). When PIPELINE_DONE = true (a docs/alm/last-pipeline.json exists), compare each chosen target stage in PP_STAGES against that file's stages[] (match by targetEnvironmentUrl origin, then by stage name). For every chosen target that has no matching stage on the existing pipeline, record a { type: "warning" } entry for the plan's Risks section: "Chosen target '{label}' ({envUrl}) has no matching stage on the existing pipeline '{pipelineName}' — setup-pipeline will need to add it." This catches the real mismatch where a saved plan says Dev → Production directly but the live pipeline only has a single Deploy to Staging stage. It is informational only — do not block or re-prompt; the user can still approve the plan, and setup-pipeline reconciles the actual stages at execution time.
🚦 Gate (plan · plan-alm:2.q4-host): Host environment selection — branches on
HOST_RESOLUTION.statusand surfaces the right menu (use-detected / pick from list / NoHost host-type / Sandbox confirm / CannotRedirect block / manual paste). DrivesHOST_ENV_URLandWILL_PROVISION_*flags for the rest of plan-alm and ensure-pipelines-host. UsesAskUserQuestionper branch.Trigger: Phase 2 Q4 entry;
HOST_RESOLUTIONpopulated in Phase 1 step 12. Why we ask: Auto-picking a host can provision a new Custom Host (WILL_PROVISION_CUSTOM) consuming an Azure capacity quota the user didn't intend; or pick the wrong env, sending pipelines through a foreign host. The downstream ensure-pipelines-host skill TRUSTS this answer and skips its own 3.C menu. Cancel leaves: Nothing — no provisioning fired yet.
Q4 (host environment — branches on HOST_RESOLUTION.status from Phase 1 step 12):
This question consumes HOST_RESOLUTION populated by the new detect-only wrapper run in Phase 1 step 12. Each branch sets HOST_ENV_URL (which feeds the rest of plan-alm) and may also set the auxiliary flags CHOSEN_ENV_URL, WILL_PROVISION_PLATFORM, WILL_PROVISION_CUSTOM, WILL_USE_PPAC, WILL_ENSURE_HOST, and USER_CHOSE_DEFER_TO_SETUP_PIPELINE. Defaults: HOST_ENV_URL = HOST_RESOLUTION.finalHostEnvUrl, all flags false / null.
Why the NoHost branch presents the env-first menu here instead of deferring to ensure-pipelines-host Phase 3.C: the original design asked a yes/no "we'll provision new — continue?" question in plan-alm and let 3.C surface the env-first choice at execution time. In practice the agent treated the plan-alm yes-confirmation as authorization to skip 3.C entirely (or to skip 4.A's pre-call gate), and users hit 4.A → 409 trial-license errors when an existing env install (4.B) would have been a clean path. Surfacing the env-first menu here — once, at planning time, when the user has full context — eliminates the ambiguity. ensure-pipelines-host then trusts CHOSEN_ENV_URL and skips its own 3.C menu (see ensure-pipelines-host Phase 3 skip rule).
status | Q4 prompt | Result |
|---|---|---|
AvailableUsingCustomHost, AvailableUsingCustomHostByAdminDefault, AvailableUsingPlatformHost | "Detected host {finalHostEnvUrl} (Pipelines v{pipelinesSolutionVersion}). Use this host?" Options: 1. Yes, use this / 2. Use a different host environment (Other) | Y → HOST_ENV_URL = HOST_RESOLUTION.finalHostEnvUrl. N → fall back to today's "enter different URL" branch (free-text via Other). |
AvailableUnboundCustomHost | "Existing Custom Host {displayName} ({finalHostEnvUrl}) found in tenant — not yet bound to dev env. setup-pipeline will reuse it (recommended; avoids duplicates). Use this host?" Options: 1. Yes, use this / 2. Use a different host environment (Other) | Y → HOST_ENV_URL = HOST_RESOLUTION.finalHostEnvUrl, WILL_ENSURE_HOST = true. N → fall back to "enter different URL". |
MultipleUnboundCustomHosts | "{N} Custom Hosts found in tenant. Which one should setup-pipeline use?" Options: enumerate HOST_RESOLUTION.candidates.existingCustomHosts[] (up to 3) by display name + URL, plus "Other" for a custom URL, plus "Decide later — setup-pipeline will ask". | Picked candidate → HOST_ENV_URL = candidate.instanceApiUrl, WILL_ENSURE_HOST = true. Decide-later → HOST_ENV_URL = null, WILL_ENSURE_HOST = true, USER_CHOSE_DEFER_TO_SETUP_PIPELINE = true. |
PlatformHostExistsUnbound | "Tenant Platform Host {finalHostEnvUrl} exists. Use it (no admin role required) or create a new Custom Host?" Options: 1. Use Platform Host / 2. Create new Custom Host / 3. Cancel | 1 → HOST_ENV_URL = HOST_RESOLUTION.finalHostEnvUrl, WILL_ENSURE_HOST = true. 2 → HOST_ENV_URL = null, WILL_PROVISION_CUSTOM = true, WILL_ENSURE_HOST = true. 3 → exit. |
NoHost | Host-type prompt — same shape as ensure-pipelines-host Phase 3.C so the user makes the host choice once, here, instead of being asked again at execution time. Present: "No Pipelines host bound to {devEnvUrl}. Which environment should host Pipelines? Pipelines lives in one env per tenant; pipelines, stages, and run history are stored there. Source envs deploy through it." Top-level options: 1. "Provision a Platform Host (recommended) — Microsoft-managed Dataverse env auto-provisioned in your tenant home geo. Pipelines app pre-installed. Idempotent. ~3–5 min." 2. "Set up a Custom Host — Pipelines lives in a Dataverse env you control. We'll ask whether to use an existing env or create a brand-new dedicated one." 3. "Open PPAC and create one manually (admin fallback)." 4. "Switch to Manual export/import strategy." 5. "Cancel." When the user picks Option 2, surface the Custom-Host sub-prompt: build the eligible-env list from HOST_RESOLUTION.candidates.eligibleForAppInstall[] with role labels (dev env, source env, staging env, production env) per origin match; cap the visible list at 5 envs with role-aware ranking (see "Eligible-env presentation cap" below). Sub-options: a. Each visible env (display name + URL + role labels) labeled "Install Pipelines app on this env" — sandbox-sku envs add a "(Sandbox — confirmation gate)" suffix; append "Other (paste URL)" as the last per-env entry. b. "Create a brand-new dedicated env (D365_ProjectHost template, ~5–10 min, requires Power Platform admin)." c. "Back — return to host-type menu." When the eligible list is empty, drop sub-option a and present only b / c. | Option 1 (Platform Host) → HOST_ENV_URL = null, WILL_PROVISION_PLATFORM = true, WILL_ENSURE_HOST = true. Option 2 → sub-prompt; sub-option a picked env → HOST_ENV_URL = picked.instanceApiUrl, CHOSEN_ENV_URL = picked.instanceApiUrl, WILL_ENSURE_HOST = true (Sandbox confirmation gate, if applicable, must be passed before this resolution stands; "Other (paste URL)" → ask for the env URL via free-text, then proceed as a picked eligible env); sub-option b → HOST_ENV_URL = null, WILL_PROVISION_CUSTOM = true, WILL_ENSURE_HOST = true; sub-option c → re-show top-level menu. Option 3 (PPAC) → HOST_ENV_URL = null, WILL_USE_PPAC = true, WILL_ENSURE_HOST = true. Option 4 (Manual strategy) → restart Phase 2 with STRATEGY = manual. Option 5 → exit. |
CannotRedirect | Block. Show the org-setting vs tenant-default mismatch error from HOST_RESOLUTION.candidates/warnings and stop the skill — only a Power Platform admin can resolve. | Exit with the specific error. |
OrgSettingStale, PermissionDenied, DetectionFailed | Surface the error; ask the user to enter the host URL manually with pac env list pre-fill (today's fallback). Pre-fill options from ENV_LIST (up to 3 known environment URLs) plus "Other" for a custom URL; pre-fill first option from docs/alm/last-pipeline.json if present. | HOST_ENV_URL = user-supplied. |
Store the resulting HOST_ENV_URL for use by the rest of plan-alm. The auxiliary flags CHOSEN_ENV_URL, WILL_PROVISION_PLATFORM, WILL_PROVISION_CUSTOM, WILL_USE_PPAC, WILL_ENSURE_HOST, and USER_CHOSE_DEFER_TO_SETUP_PIPELINE feed the planData hostResolution block in Phase 3 and the inline summary in Phase 4. ensure-pipelines-host reads chosenEnvUrl, willProvisionPlatform, willProvisionCustom, and willUsePpac from that block to bypass its own Phase 3.C menu when the user has already made the choice here.
Eligible-env presentation cap (used by the NoHost row above and by MultipleUnboundCustomHosts). When the eligible list runs long, build the visible options as follows so the prompt stays scannable:
dev env, source env, staging env, or production env label (matched by URL origin against devEnvUrl and against PP_STAGES[].envUrl). These are the project's own envs and are nearly always the right pick. Dedupe by origin.list-tenant-envs.js (name-hint pattern pipeline|deploy|host|alm|cicd|govern → admin-perms → recency).eligible.length > 5, suffix option 1's headline with: Showing top 5 of {N}; the remaining {N-5} eligible env(s) can be reached via the "Other (paste URL)" entry. When eligible.length <= 5, no suffix (all envs visible inline).ENV_LIST (the list-environments.js output gathered in Phase 1 — each entry's environmentUrl) so they can paste-or-pick from the full inventory rather than typing a URL by hand.The same cap policy applies to ensure-pipelines-host Phase 3.C — see that skill's Step 3a for the same rules. Keep the two implementations consistent so users see the same prompt shape regardless of whether they enter via plan-alm or directly via setup-pipeline → ensure-pipelines-host.
🚦 Gate (plan · plan-alm:2.q5-approval): Pick approval mode — required per stage / staging auto + prod required / no gates.
Q5: Ask via AskUserQuestion:
"Should deployments require approval before each stage?"
Options:
Store as PP_APPROVAL_MODE.
Note: PP Pipelines always exports as a managed solution to target environments. Set EXPORT_TYPE = "managed" automatically — no question needed.
Q6 (auto-detect, no question): Resolve HAS_ENV_VARS in this order:
ENV_VARS_DETAILS.length > 0 (Phase 1 Step 10b returned per-variable rows): HAS_ENV_VARS = true. This is the most accurate signal — we've enumerated the definitions directly.SPLIT_PLAN.sizeAnalysis.envVarCount.value > 0 (the size estimator counted definitions but the enumerator didn't return rows — usually because the publisher prefix or token was missing): HAS_ENV_VARS = true. The plan's count-summary fallback will explain to the user that per-variable details will be enumerated later.SOLUTION_DONE = true and .solution-manifest.json lists components with componenttype 380: HAS_ENV_VARS = true. (Stale-manifest fallback — should rarely fire now that 10b is in place.)HAS_ENV_VARS = false. Variables will be discovered during setup-solution.When HAS_ENV_VARS = true, the plan notes that deploy-pipeline will prompt for per-stage env var values (see Phase 3 risks population).
🚦 Gate (plan · plan-alm:2.q3-manual): Pick how many target environments for manual export/import path.
Q3: Ask via AskUserQuestion:
"How many target environments do you need to deploy to?"
Options:
Store as MANUAL_TARGET_COUNT.
If option 3: set MANUAL_TARGET_COUNT = 0. Proceed to Q5.
🚦 Gate (plan · plan-alm:2.q4-manual-target): Pick the URL for each manual target env. Fires PER TARGET in the
MANUAL_TARGET_COUNTloop. Two targets (Staging + Production) = two separate prompts. Each prompt pre-fills frompac env listand accepts a different URL. Do NOT collect all target URLs in a single multi-input prompt — each target is a distinct decision (different audiences, different env characteristics, possibly different SKUs).
Q4 (one per stage): For each target environment needed, ask via AskUserQuestion:
"What is the URL for target environment {N}?"
Pre-fill from ENV_LIST: show up to 3 known environment URLs from pac env list as options, plus "Enter a different URL" as the last option.
Store target URLs as MANUAL_TARGETS (array).
🚦 Gate (consent · plan-alm:2.q5-manual-type): Managed vs Unmanaged export — irreversible choice for the produced zip.
Q5: Ask via AskUserQuestion:
"How should the solution be exported?"
Options:
Store as EXPORT_TYPE, and log the decision ({ field: "exportType", value: EXPORT_TYPE, source: "default"|"explicit" }).
No checkpoint preference is collected.
plan-almno longer runs the export/import itself, so the old "pause between export and import?" question (which only gated execution) is gone. The plan records the recommended sequence —export-solutionthenimport-solutionper target — and the user reviews the zip between those steps when they run them. A standing note to that effect is added to the plan's risks/steps in Phase 3.
Q6 (auto-detect, no question): Same as PP Pipelines Q6 — check for env var definitions.
Now create the two planner tasks — strategy is known. plan-alm is a planner, so it has exactly two tasks regardless of path:
| # | Subject | activeForm | Description |
|---|---|---|---|
| 1 | Generate ALM plan | Generating ALM plan | Build planData, render docs/alm-plan.html |
| 2 | Approve & save ALM plan | Awaiting plan approval | Present inline summary, capture approver, save (approved or draft) |
Do not create Setup/Deploy/Activate/Test/Import/Finalize tasks. Those are execution steps performed by the individual ALM skills the user runs after approval — they are not
plan-almtasks. The recommended sequence still appears in the plan: the planDatasteps[]array below lists it (allstatus: "pending"), and the rendered HTML's Execution Checklist shows it as the plan of record. Each downstream skill marks its own step complete (viarefresh-alm-plan-data.js) when the user runs it.
Mark task 1 ("Generate ALM plan") as in_progress.
Build a planData object with all gathered strategy inputs:
solutionContents is populated from SOLUTION_CONTENTS_DATA built in Phase 1. If discovery was unavailable, pass null — the renderer will show a fallback note.
plannedEnvVarCount is computed from SITE_SETTINGS_DATA (Phase 1 Step 7): (SITE_SETTINGS_DATA.promoteToEnvVar?.length || 0) + (SITE_SETTINGS_DATA.credentialNeedsDecision?.length || 0). When SITE_SETTINGS_DATA is null (Step 7 query failed), set plannedEnvVarCount = 0. The renderer reads this alongside envVars.length (existing) and sizeAnalysis.envVarCount.value (size-estimator's count) to produce a "N today / +M planned" display in the Overview stat card and Size Analysis signal.
validationRuns block (PP Pipelines path only — initialize one entry per target stage with value null). plan-alm leaves these null at plan time; they are populated later by test-site's own final-phase refresh (refresh-alm-plan-data.js --phase test-site --stageName {stage}) when the user runs test-site for each stage. The full categorized test report drives the Validation tab in the rendered HTML. Shape per stage:
The shape is identical to docs/alm/last-test-site.json written by test-site Phase 6.7a — the refresh helper reads that file verbatim and assigns it to validationRuns[stageName] when test-site runs. The renderer maps runOutcome to green / yellow / red Outcome badges and produces a sub-tab per stage on the Validation tab. For the Manual path, omit validationRuns from planData.
pipelineMeta block (PP Pipelines path only — read from docs/alm/last-pipeline.json and docs/alm/last-deploy.json at planData-build time. null on fresh plans where no pipeline is configured yet. Later refreshed by setup-pipeline's and deploy-pipeline's own final-phase refresh when the user runs them). Highlights the pipeline that is actually moving configurations for this project. Shape:
isActive: true whenever the project has a configured pipeline (docs/alm/last-pipeline.json exists). Drives the ACTIVE chip on the Pipelines tab.pipelineName: from docs/alm/last-pipeline.json. The renderer falls back to ${SITE_NAME}-Pipeline when pipelineMeta is absent.reusedByWiring: null when the pipeline was created fresh; an object { originalName, requestedName } when create-deployment-pipeline.js matched an existing pipeline by source+target wiring and reused it under its existing name. The renderer surfaces this with an explanatory note so reviewers understand why the plan and the live pipeline names may differ.lastDeploy: derived from docs/alm/last-deploy.json. Omit (set to null) before the first deploy.How to populate. During Phase 3 planData build, read both files (Node.js inline) and inject:
Embed the result as planData.pipelineMeta.
v2 fields (sizeAnalysis, assetAdvisory, splitStrategy, appliedStrategies, compositeSubPartitioned, proposedSolutions, recommendations, breakdown) come straight from SPLIT_PLAN computed in Phase 1 Step 10, mutated by Q1b user choices. Pass them through unchanged to the renderer. envVars comes from ENV_VARS_DETAILS populated in Phase 1 Step 10b — pass it through unchanged. When the array is empty, the renderer's count-aware fallback uses sizeAnalysis.envVarCount.value so the Env Variables tab still shows a count-summary note.
Disk-measurement cross-check (webFilesDiskMeasuredMB, webFilesDiskMeasuredPath, webFileSampleSize) — hoist these from the estimator output (rawDiscovery.estimate.webFilesDiskMeasuredMB, etc.) onto the top level of planData. They're only populated when --projectRoot "." was passed to estimate-solution-size.js AND a build-output directory was detected. The renderer surfaces webFilesDiskMeasuredMB next to the Dataverse-measured Web Files signal when the two numbers disagree materially (the same condition that flips truncationSuspected), so reviewers can see at a glance which number to trust. Pass null for all three when the estimator didn't measure disk.
Truncation canary (truncationSuspected, truncationWarnings) — pass through verbatim from SPLIT_PLAN.truncationSuspected / SPLIT_PLAN.truncationWarnings (which compute-split-plan.js itself reads from the estimate). When truncationSuspected === true, the rendered plan shows a red banner above the size analysis, and the Phase 2 Q1b "keep as single anyway" override is gated by an extra confirmation step (see Phase 2 Q1b). Common causes: Dataverse pagination regression in estimate-solution-size.js OR a web-file undercount (file-typed columns whose bytes aren't returned via $select=content) — the disk cross-check (when enabled) flags the latter explicitly.
rawDiscovery block — embed the full estimator + split-plan + host-resolution outputs verbatim, before any field mapping. This is a diagnostic envelope: in general, the renderer must not reference its contents from display paths. Narrow exception: the renderer's webFilesDiskMeasured* / webFileSampleSize read paths fall back to rawDiscovery.estimate.* if the top-level hoist is missing, so older plan files (written before the hoist was specified) still render the disk-compare note. The right fix when you see that fallback fire is to re-build planData with the hoist populated, not to keep the fallback path active. Otherwise the principle stands: rawDiscovery exists so a reviewer (or a future agent debugging a wrong plan) can read docs/.alm-plan-data.json and see what the discovery scripts actually produced, side-by-side with the mapped fields the renderer used. Read each source file once and assign:
Use null for any source that was skipped (Manual path doesn't run host resolution; pre-pipeline projects don't have a manifest). Do not redact or summarize — the value of the snapshot is that it's the raw input.
proposedSolutions[]is never empty. Even whensplitStrategy === "single"and the decision tree recommends one solution,compute-split-plan.jsreturns one entry describing the base solution (uniqueName, displayName, sizeMB, componentCount, componentTypes). Pass that through. The renderer drives the Solutions tab off this array — leaving it empty produces an unhelpful "structure will be determined" placeholder. If you find yourself withproposedSolutions = []because compute-split-plan wasn't run, synthesize a single base entry fromsolutionContents.solution/data.SITE_NAME/componentCount/totalSizeMBrather than passing through an empty array. The renderer has a safety-net synthesizer for this case but the right fix is upstream — populate it explicitly.
hostResolution block (PP Pipelines path only — omit for Manual path). Built from HOST_RESOLUTION (Phase 1 step 12) plus the auxiliary flags set by Phase 2 Q4:
status ← HOST_RESOLUTION.statushostEnvUrl ← HOST_ENV_URL (from Q4) — may be null when the user deferred or chose to provision newhostEnvId ← HOST_RESOLUTION.finalHostEnvIdhostType ← HOST_RESOLUTION.hostTypepipelinesSolutionVersion ← HOST_RESOLUTION.pipelinesSolutionVersioncandidatesCount ← HOST_RESOLUTION.candidates.existingCustomHosts.lengthwillEnsureDuringExecution ← WILL_ENSURE_HOST flag from Q4 (true whenever setup-pipeline will need to consult ensure-pipelines-host at execution time — i.e. status is NoHost, any *Unbound*, or the user deferred)willProvisionPlatform ← WILL_PROVISION_PLATFORM flag from Q4 (set when the user picks Option 1 "Provision a Platform Host" in the NoHost host-type menu; ensure-pipelines-host treats this as a directive to go straight to Phase 4.0)willProvisionCustom ← WILL_PROVISION_CUSTOM flag from Q4willUsePpac ← WILL_USE_PPAC flag from Q4 (set when the user picks "Open PPAC and create one manually" in the NoHost host-type menu; ensure-pipelines-host treats this as a directive to go straight to Phase 4.C)chosenEnvUrl ← CHOSEN_ENV_URL flag from Q4 (set when the user picks Option 2 → sub-option a in the NoHost host-type menu; ensure-pipelines-host treats this as a directive to skip its Phase 3.C menu and go straight to Phase 4.B with this env)userChoseDeferToSetupPipeline ← USER_CHOSE_DEFER_TO_SETUP_PIPELINE flag from Q4 (only set in the MultipleUnboundCustomHosts "Decide later" branch)Populate risks based on gathered data:
HAS_ENV_VARS = true: { type: "warning", message: "This solution has environment variables ({N} detected) — you will be prompted for per-stage values during deployment." }. Substitute {N} from ENV_VARS_DETAILS.length if it's > 0, otherwise from SPLIT_PLAN.sizeAnalysis.envVarCount.value (the count the size estimator reported). When neither source has a positive count, drop the parenthetical ("This solution has environment variables — you will be prompted...").SITE_SETTINGS_DATA.promoteToEnvVar.length > 0: { type: "info", message: "{N} auth-related site settings (Authentication/* and AzureAD/*) detected with values. setup-solution will ask which to back with environment variables so each stage can use different values (e.g., different OAuth callback URLs). Skip any you don't need to vary per stage." }. Substitute {N} from SITE_SETTINGS_DATA.promoteToEnvVar.length. (Replaces older "will be promoted" wording — that implied automatic action; in reality the user picks per setting.)SITE_SETTINGS_DATA.credentialNeedsDecision.length > 0: { type: "info", message: "{N} credential-style site settings (ConsumerKey / ClientId / ClientSecret / etc.) detected. setup-solution will run a single bulk prompt to handle all of them — auto-classify by name (recommended; *Secret/Password/ApiKey/AppKey become Secret env vars, *Id/ConsumerKey become String env vars), all-as-Secret, all-as-String, skip-all, or fall through to a per-credential picker for granular control." }. Substitute {N} from SITE_SETTINGS_DATA.credentialNeedsDecision.length. Do NOT emit any "excluded from solution / configure manually" wording — that was the pre-IronItOut behavior and it's gone.EXPORT_TYPE = "unmanaged" and strategy includes a production target: { type: "warning", message: "Unmanaged solutions can be edited in the target environment — consider using Managed for production." }SOLUTION_DONE = false: { type: "info", message: "A Dataverse solution will be created first — publisher prefix is irreversible once chosen." }SOLUTION_DONE becoming true after Phase 4): { type: "info", message: "When you add new components later (server logic, cloud flows, env vars, custom tables), re-run /power-pages:setup-solution in sync mode to bring them into this solution. The completeness check in this skill (Phase 1 Step 11) will flag any drift between the live site and the solution before the next plan-alm run." }. Skip when SOLUTION_DONE is already true at plan-alm start (sync mode is already self-evident at that point).KNOWN_GAPS is set (the pre-plan completeness check in Phase 1 Step 11 found gaps and the user chose to continue): { type: "warning", message: "{X} site components, {Y} cloud flows, {Z} env vars, and {W} custom tables exist on the site but are not in the current solution. This plan will not promote them — run /power-pages:setup-solution sync mode before deploying, or re-run plan-alm after syncing." }. Substitute the counts from KNOWN_GAPS.missing.*.length.HOST_RESOLUTION.status === "NoHost" AND WILL_PROVISION_PLATFORM === true: { type: "info", message: "No Pipelines host detected. setup-pipeline will provision a new Platform Host (idempotent, ~3–5 min). Plan execution will pause for a tenant-identity confirmation gate before the call." }. Do NOT include any wording about admin-role requirements or API names — those are implementation details.HOST_RESOLUTION.status === "NoHost" AND WILL_PROVISION_CUSTOM === true: { type: "info", message: "No Pipelines host detected. setup-pipeline will create a new Custom Host. Plan execution will pause for admin-role attestation and a pre-call confirmation." }HOST_RESOLUTION.status === "NoHost" AND none of the provisioning flags are set (user picked an existing env via Option 2 → sub-option a, or chose PPAC manual, or deferred): emit a status-appropriate info entry derived from CHOSEN_ENV_URL / WILL_USE_PPAC.HOST_RESOLUTION.status === "AvailableUnboundCustomHost": { type: "info", message: "An existing Custom Host (" + HOST_RESOLUTION.finalHostEnvUrl + ") will be reused. Source env will be bound to it automatically." }HOST_RESOLUTION.status === "MultipleUnboundCustomHosts": { type: "info", message: HOST_RESOLUTION.candidates.existingCustomHosts.length + " existing Custom Hosts found in tenant. setup-pipeline will prompt for selection." }HOST_RESOLUTION.status === "PlatformHostExistsUnbound": { type: "info", message: "Tenant has a Platform Host. Reusing it is the lowest-friction option; creating a Custom Host instead provides better governance for separate-tenant or governed scenarios." }HOST_RESOLUTION.status === "CannotRedirect": { type: "warning", message: "CannotRedirect: source env ProjectHostEnvironmentId points at PE but tenant default custom host is set elsewhere. Resolution requires Power Platform admin." } (Note: Phase 2 Q4 normally blocks plan generation in this state; this is a defensive entry in case the plan is somehow generated.)STRATEGY = "manual"): { type: "info", message: "Recommended sequence: run /power-pages:export-solution, review the produced zip, then run /power-pages:import-solution for each target. plan-alm does not perform the export/import itself." }rawDiscovery.estimate, rawDiscovery.splitPlan, and (PP path) rawDiscovery.hostResolution that is null at planData-build time: { type: "warning", message: "Discovery for {X} did not run; the related size/split/host decisions in this plan are unverified." } (substitute {X} = "solution size estimate" / "split analysis" / "pipeline host resolution"). Carve-out for rawDiscovery.hostResolution: when PIPELINE_DONE = true, host resolution is intentionally skipped (Phase 1 Step 12's skip rule — the host comes from docs/alm/last-pipeline.json, not a fresh probe), so a null hostResolution is expected, not a gap. Do NOT emit the "pipeline host resolution" warning when PIPELINE_DONE = true — it would be a spurious "host resolution did not run" on every project that already has a pipeline. The estimate and splitPlan gap warnings still apply regardless of PIPELINE_DONE.Plan completeness check (#10). Before writing planData, verify the plan rests on real discovery:
sizeAnalysis.totalSizeMB is non-null,stages[] is non-empty,solutionContents is populated when a solution is configured (SOLUTION_DONE = true).If any check fails, set PLAN_QUALITY = "degraded" and record which check failed. (This is in addition to the Phase 1 Step 6 auth-failure path that already sets degraded.)
Degraded-plan risk (#8) — if PLAN_QUALITY === "degraded" (from the auth-failure path or the completeness check), prepend to risks: { type: "warning", message: "⚠ PLAN QUALITY: DEGRADED — discovery was incomplete ({cause}). Review carefully before executing; re-run /power-pages:plan-alm after fixing the cause (commonly dev-environment auth) to regenerate a complete plan." }. Use type: "warning" (the renderer styles it). Substitute {cause} from the recorded reason(s).
Write planData to docs/.alm-plan-data.json (create docs/ if it doesn't exist).
Keep docs/.alm-plan-data.json on disk — never delete it. Two consumers depend on it after plan-alm exits: (1) the execution skills' final-phase refresh (refresh-alm-plan-data.js) reads it to update hostResolution, pipelineMeta, validationRuns, risks, steps[] status, and the footer as each skill runs, then re-renders docs/alm-plan.html so the rendered tabs reflect actual run state (not the pre-run plan); (2) check-alm-plan.js reads it for the Phase 0 ALM-plan gate in setup-pipeline / deploy-pipeline / setup-solution / export-solution / import-solution / configure-env-variables — deleting it makes every downstream skill think no plan exists (and fire its no-plan gate). Earlier guidance to delete this file after the initial render was incorrect and caused the Pipelines tab + risks list to stay frozen at pre-run state for the lifetime of the plan.
Write docs/alm/alm-plan-context.json (persists so setup-solution can read it):
This file is intentionally NOT deleted — setup-solution and other skills read it to skip re-discovery.
The inline Markdown summary presented in Phase 4 is intentionally compact — reviewers need to see the full rendered plan (size gauge, signal cards, per-solution breakdown, asset advisory, pipeline stages) before giving informed approval. Launch docs/alm-plan.html in the default browser before the approval prompt so the user can scan the full plan while reading the CLI summary.
Why no Node wrapper. The earlier
node -e "spawn('powershell.exe', [...])"chain hits the agent's sandbox classifier (Node spawning a process that spawns another process is the textbook pattern the classifier blocks). The agent should use the OS-native shell tool directly — no Node detour, nochild_process, no detached subprocess.
Step 1. Print the absolute file:// URL prominently first. This is the user's reliable fallback if the launcher gets blocked:
(Single Node call, no spawn, never blocked. Output: Plan URL: file:///C:/Projects/.../docs/alm-plan.html.)
Step 2. Launch the browser via the OS-native shell. Pick the shell tool the agent has available:
Windows / PowerShell tool — call Start-Process directly:
Windows / Bash tool (Git Bash, WSL passthrough) — call PowerShell from Bash, but as a single direct invocation (no Node wrapper):
macOS — call open directly:
Linux — call xdg-open directly:
Step 3. Report the URL to the user. After the launch attempt, surface the file:// URL the agent printed in Step 1, so the user always has a clickable backup:
"Opened
docs/alm-plan.htmlin your browser. If it didn't open automatically, use this link:file:///C:/Projects/.../docs/alm-plan.html. Review it, then answer the approval prompt below."
If the launch silently fails (sandboxed terminal, SSH session, headless environment), do not retry, do not block, do not loop. Step 1's printed URL is the contract — the user can click or paste it themselves. Continue to Phase 4 and rely on the CLI summary as backup.
Mark task 1 as completed.
Mark task 2 ("Approve & save ALM plan") as in_progress.
plan-alm stops here. This phase saves the plan (approved or draft) and exits. It does not invoke setup-solution, setup-pipeline, deploy-pipeline, export-solution, or import-solution. The user runs those afterward.
Present a concise inline Markdown summary:
🚦 Gate (plan · plan-alm:4.approve): The capstone — user saves the rendered HTML plan.
plan-almnever executes; this gate only chooses whether the plan is saved Approved (ready for the user to run the execution skills) or Draft, or sends the user back to revise. No downstream skill is dispatched here.
Ask via AskUserQuestion:
"How would you like to save this ALM plan?"
Options:
If option 3: Re-run Phase 2 (ask which section to change, then re-gather those answers). Regenerate the plan (repeat Phase 3). Re-present for approval.
If option 1 (approved): Capture the approver (see below), then write the status and approver atomically with the deterministic helper. Do not hand-edit the HTML spans — the badge, approved-by, and approval-date are all derived from docs/.alm-plan-data.json on render, so a manual span Edit is non-durable (the next execution-skill refresh re-derives it). The helper updates plan-data (PLAN_STATUS, PLAN_MODE, APPROVED_BY, APPROVAL_DATE — all four together) and re-renders docs/alm-plan.html:
Then run the finalize steps below (skill tracking + commit), print the next-steps guidance, mark task 2 completed, and exit.
If option 2 (draft): Do not capture an approver. Save as Draft with the same helper (writes plan-data + re-renders; clears any stale approver fields so the plan never ends up "Draft + approver"):
Run the finalize steps below (commit only — skip skill tracking or run it, your choice; commit message "Add ALM plan for {siteName} (draft)"), tell the user to re-run /power-pages:plan-alm when ready to approve, mark task 2 completed, and exit.
Capturing the approver (option 1 only) — always interactive (#1):
Never auto-apply a name silently. First compute a prefill suggestion from git / OS user:
Then always ask via AskUserQuestion (even when the suggestion is non-empty) so a human actively confirms the approver recorded in the audit trail:
"Who is approving this plan? (recorded in the audit trail in docs/alm-plan.html)"
Options: 1. {suggested name from git/OS, if any} · 2. Other (enter name)
If the command returned an empty string, present only option 2 (free-text). Store the confirmed result as APPROVER and pass it to set-plan-status.js (option 1 above) — do not hand-edit the approved-by / approval-date spans. The helper writes APPROVED_BY + APPROVAL_DATE (current ISO timestamp) into docs/.alm-plan-data.json and re-renders, and the template fills both spans from plan-data. This keeps the audit trail and the status in lockstep — the half-written "approver recorded but status still Draft" state (which validate-plan-alm.js now blocks) cannot happen.
Finalize (both save options):
Reference:
${PLUGIN_ROOT}/references/skill-tracking-reference.md
docs/.alm-plan-data.json stays on disk — it is read by check-alm-plan.js for every downstream skill's Phase 0 gate and refreshed by those skills as they run. Never delete it.Next-steps guidance (option 1 — print to the user, do NOT invoke):
"Plan approved and saved. plan-alm doesn't deploy — run these next, in order. Each detects this plan and proceeds without re-asking, then updates the plan as it completes:
- PP Pipelines path:
/power-pages:setup-solution→/power-pages:setup-pipeline→/power-pages:deploy-pipeline(once per target stage; the deploy flow activates the site for you — but if you defer activation, run/power-pages:activate-site) →/power-pages:test-siteto validate each stage.- Manual path:
/power-pages:setup-solution→/power-pages:export-solution→ review the zip →/power-pages:import-solution(once per target). Skip any step marked already set up. You can re-open the plan any time atdocs/alm-plan.html."
| Task subject | activeForm | Description |
|---|---|---|
| Generate ALM plan | Generating ALM plan | Gather strategy inputs, build planData, render docs/alm-plan.html |
| Approve & save ALM plan | Awaiting plan approval | Present inline summary (incl. defaulted-decisions + plan quality), capture approver, save Approved or Draft, commit, print next-steps |
plan-almhas exactly these two tasks. Setup / pipeline / deploy / export / import / activate / test are performed by the individual ALM skills the user runs after approval — they appear in the plan'ssteps[](the recommended sequence) but are neverplan-almtasks.
.alm-deferred marker handling; pre-plan completeness check (if a solution exists)setup-solution in plan.powerpages-site/website.yml and no powerpages.config.json: stop, advise /power-pages:create-site (Phase 1 resolves site identity from either marker — a data-model/EDM site has only website.yml, so don't hard-stop on the missing config alone)pac env list fails: skip ENV_LIST pre-filling; ask for environment URLs manuallyrender-alm-plan.js fails (non-zero exit): report error, show planData JSON as fallback, ask user whether to proceedPLAN_QUALITY = "degraded", surface a prominent risk, still produce the plan (the user fixes auth and re-runs to regenerate)