Prompt detail

Metric Discrepancy Check

Verification of a difference between a DevStats number and another card, external source, or user expectation.

Prompt 12
Analyzed reports
PR Cycle TimeIssue Cycle TimeThroughputDeployDORA MetricsCode ReviewSprintsPlanning AccuracyPR ListIssue List
Context to provide

Provide the metric, Squad-Name scope, period, DevStats value, expected value, and the source or card being compared. Include known filters, date fields, units, and rounding rules.

Copyable prompt
2,528 characters
ROLE Engineering data investigator using MCP DevStats. OBJECTIVE Investigate why a DevStats metric value appears to diverge from an expected value, external source, card, or user perception. Classify the discrepancy by scope, period, filter, definition, coverage, precision, or unresolved limit. Do not assert a data error without evidence and do not invent details from a missing source. ...

How the analysis runs

The check restates both values, aligns definitions and query boundaries, and classifies the discrepancy as explained or unresolved without assuming a product defect.

  1. 01Scope and inputsDocument the two values being comparedCapture Squad-Name, metric definitions, dates, filters, units, aggregation, rounding, and source details before running a reconciliation.View detailsHide details
    Capture Squad-Name, metric definitions, dates, filters, units, aggregation, rounding, and source details before running a reconciliation.
  2. 02Analysis and evidenceReconcile population and calculationRun the matching DevStats report and use row-level evidence only when it can test a suspected scope, filter, date-field, or denominator difference.View detailsHide details
    Run the matching DevStats report and use row-level evidence only when it can test a suspected scope, filter, date-field, or denominator difference.
  3. 03Interpretation limitsClassify or delimit the discrepancyReport whether the gap comes from scope, period, filter, definition, coverage, or precision; otherwise state exactly what remains unknown.View detailsHide details
    Report whether the gap comes from scope, period, filter, definition, coverage, or precision; otherwise state exactly what remains unknown.

Expected Output

Markdown · Illustrative example

Metric Discrepancy Check — Throughput — Squad-Name — Jun 2026

Values compared

SourceValueKnown definition
DevStats Throughput42 merged PRsmerged_at within 1–30 Jun, all mapped repositories
External dashboard47 PRscreated_at within 1–30 Jun, two repositories

Classification

Definition and scope mismatch.

Evidence

  • Using merged_at and all mapped repositories returns 42 PRs.
  • Filtering DevStats to the two external repositories returns 39 PRs.
  • Counting PRs created in June for those repositories returns 47, matching the external value.
  • Eight PRs were created in June but merged outside the month.

What remains unknown

  • The external dashboard's timezone and bot-account exclusion are undocumented.

Validation questions

  • Does the external source intentionally measure created rather than merged PRs?
  • Which timezone and account exclusions does it apply?

Evidence used

  • Throughput — reproduce the DevStats value
  • PR List — validate date field and repository population

Stop guessing. Start asking.

Connect DevStats MCP to your AI assistant and turn these prompts into instant answers. Copy a prompt and go.

Get a demoStart free trial