analytics · microsoft — Claude Code skill: Process Mining routing and overview
analytics is a Claude Code skill that routes Process Mining queries, shows process and event/variant metrics, and loads a specialist skill for deeper analysis.
MIT ★ 1,167 updated 2026-10-29
/plugin marketplace add microsoft/power-platform-skills
/plugin install process-intelligence@power-platform-skillsWhat analytics does
Skill reads the process list, picks one and fetches base metrics: event count, case count, variants, mean duration and waiting time. When the user asks for path concentration, the skill returns the five most frequent variants. After the base summary, the skill analyses the question and routes to a specialist: investigate-process for hypotheses, analyze-performance for bottlenecks, analyze-variants for repeated work, analyze-drivers for attributes, compare-cohorts for A/B cohorts.
Skill requires a working Power Platform Process Intelligence installation and does not create visualizations: it reads data and hands off control. An empty process list stops analysis.
What it helps with
- Launch a Process Mining overview from the IDE
- Select a process and fetch event, case and variant metrics
- Show the five most frequent execution variants
- Hand off to a specialist skill for performance or attribute analysis
- Read metrics without creating visualizations or dashboards
Example prompts for the agent
- Show an overview of the
OrderFulfillmentprocess from Process Mining - What are the case and event metrics for the
IncidentResolutionprocess? - List the five most common variants for
CustomerOnboarding
How to install analytics
1Plugin marketplace
/plugin marketplace add microsoft/power-platform-skills/plugin install process-intelligence@power-platform-skillsThis installs the whole process-intelligence plugin with all its skills
2Skills CLI (Vercel)
npx skills add microsoft/power-platform-skills --skill analytics -a claude-code3Manually: put the files in place
~/.claude/skills/analytics/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/analytics/
1Skills CLI (Vercel)
npx skills add microsoft/power-platform-skills --skill analytics -a codex2Manually: put the files in place
.agents/skills/analytics/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add microsoft/power-platform-skills --skill analytics -a cursor2Manually: put the files in place
.cursor/skills/analytics/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install microsoft/power-platform-skills plugins/process-intelligence/skills/analytics2Manually: put the files in place
.github/skills/analytics/SKILL.md1Skills CLI (Vercel)
npx skills add microsoft/power-platform-skills --skill analytics -a gemini-cli2Platform CLI
gemini skills install https://github.com/microsoft/power-platform-skills --path plugins/process-intelligence/skills/analyticsManually: put the files in place
.devin/skills/analytics/SKILL.mdThe legacy .windsurf/skills/ path still works
Claude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about analytics
How to install analytics in Claude Code?
Run /plugin marketplace add microsoft/power-platform-skills, then /plugin install process-intelligence@power-platform-skills. Via CLI: npx skills add microsoft/power-platform-skills --skill analytics -a claude-code. The skill appears in ~/.claude/skills/analytics/SKILL.md.
What does the analytics skill do?
Skill fetches base process metrics from Process Mining: event count, case count, variants, mean duration. After the overview, it routes to a specialist skill for performance, variant, attribute or A/B cohort analysis.
What does analytics require to work?
Skill requires an installed Power Platform Process Intelligence and access to processes via its API. Without processes in the system, analysis will not start.
Source text
SKILL.md · MIT
---
name: analytics
description: Use when a Process Mining overview, initial process selection, or an analytical follow-up needs routing.
allowed-tools: Read, AskUserQuestion
---
# Process overview and routing
Use the [shared contract](../../references/analysis-contract.md) once per fresh analysis context.
Recipes name native functions, not provider-prefixed host tools; bind them to discovered schemas.
## Shortest trace
1. Reuse the selected process and fresh metadata. Otherwise `get_processes {}` ->
resolve the returned `P` (ask on duplicates) -> `get_process_details_v2 {processId:P}`
when metadata is missing. An empty process list stops analysis.
2. `get_process_overall_metrics_v2 {S,filterOptions:F}`. Omit absent filters.
Report returned case/event/variant counts and mean `CaseDuration`/`CaseWaitingTime`;
preserve native units. Missing metrics are not zero.
For utilization with zero active time or start-only data, check the zero-duration convention
in [filters and units](../../references/filters-and-units.md); do not infer productive time.
3. Only for requested path concentration:
`get_variants_with_metrics_v2 {S,filterOptions:F,itemsPerPage:5,itemsToSkip:0,metricToSortBy:"CaseCount",sortOrder:"Descending"}`.
Reuse completed evidence; continue necessary pages or checks without fixed analytical-call
or total-row quotas. Stop at an answer, not an arbitrary count; do not sweep or repeat unchanged queries.
Do not force an overview or investigation loop before a specialist's exact query.
Broad business questions or competing explanations use investigate-process: establish only
missing outcome definitions that change interpretation, then test the relevant claim.
| Follow-up | Load only this skill |
|---|---|
| Broad business question, observed pattern, competing explanations | investigate-process |
| Slow activities, waiting, handoffs, cumulative workload | analyze-performance |
| Common/slow variants, repeated work, a specific case | analyze-variants |
| Explanatory attributes, association hypotheses | analyze-drivers |
| Before/after or A/B cohorts | compare-cohorts |
| Missing metric, saved metric reuse, ephemeral formula | derive-metric |
| Interacting objects or leading-object executions | analyze-objects |
Stop when answered. Return a brief finding, at most five headline rows, scope/version,
metric aggregation/units and coverage; retain compact context for the next question.
Views, visualizations and saved metrics cannot be created through these tools.
© Microsoft · MIT · microsoft/power-platform-skills
Similar skills
All skills for Data and databases →analytics by other authors comes first.