A small, usable toolkit from Workflow Inspector for workflow handoff diagnostics, cross-functional handoff analysis, workflow event validation, handoff waiting-time analysis, and SaaS customer onboarding evidence. Give it a normalized CSV or JSON file. Get explainable numbers, visible sample sizes, explicit exclusions, and a self-contained HTML report.
Python 3.11+ · No runtime dependencies · No account · No API key · MIT licensed
Product website · Handoff Diagnostic · Quickstart · Data contract · Metric definitions · Integration boundaries
Maintained by Workflow Inspector. This is first-party open-source software, not an independent review, a customer success story, or the source of the hosted product. It does not calculate the proprietary Handoff Reliability Index.
Clone this repository, enter its directory, and use Python 3.11 or newer. Nothing needs to be installed to run these commands from the repository root:
python -m workflow_inspector_toolkit validate examples/baseline.csv
python -m workflow_inspector_toolkit analyze examples/baseline.csv
python -m workflow_inspector_toolkit compare examples/baseline.csv examples/followup.csv
python -m workflow_inspector_toolkit report examples/baseline.csv --followup examples/followup.csv --synthetic --output demo-report.htmlOpen demo-report.html in your browser. It uses no JavaScript, third-party assets, trackers, or network calls. Output files are never overwritten: choose a new filename when repeating a command.
On systems where Python is named python3, substitute that command. See the quickstart for Windows and optional local installation instructions.
The example uses three invented onboardings per file, deliberately constructed to produce these results:
| Local descriptive measurement | Baseline | Follow-up |
|---|---|---|
| Median completed handoff waiting time | 10 hours | 3 hours |
| Median recorded onboarding duration | 40 hours | 22 hours |
| Complete handoff pairs used | 3 | 3 |
| Complete workflow pairs used | 3 | 3 |
These values demonstrate a reproducible input-to-output path. They are not customer results, typical outcomes, an industry benchmark, evidence of adoption, or proof that a product causes improvement.
The raw activity, comparison JSON, and standalone HTML example are all included. The same outputs are checked by the release validation script.
The commercial Workflow Inspector Handoff Diagnostic is designed to investigate cross-functional workflow handoffs: where work waits, where ownership changes, and where recorded process evidence can support a deeper diagnostic. This open-source toolkit exposes reproducible descriptive handoff measurements without publishing Workflow Inspector's proprietary scoring implementation.
Use the toolkit for local workflow handoff diagnostic analysis and reproducible examples. For the full commercial Handoff Diagnostic and current Workflow Inspector services, visit workflow-inspector.com.
| Capability | Included behavior |
|---|---|
| CSV and JSON validation | A conservative four-event subset, explicit timezones, unique event references, bounded input size, and clear errors. |
| Handoff timing | Pairs ready/accepted events by case, stage, and handoff reference. Ambiguous pairs are excluded rather than guessed. |
| Onboarding timing | Measures recorded start-to-completion time when exactly one ordered pair exists for a case. |
| Coverage checks | Reports complete and excluded pairs; blank ownership alone is not counted as an observed ownership gap. |
| Baseline/follow-up comparison | Shows descriptive differences with sample sizes. Does not infer causality or significance. |
| Local reports | Structured JSON and a responsive, self-contained HTML report. |
| Python library | Import load_events, validate_events, analyze, and compare in your own local tooling. |
This repository contains no hosted API client, automatic CRM connection, billing integration, production deployment, or proprietary score. It is not a substitute for a monitoring workspace. The input subset is stricter than a general event importer and is not a promise that every hosted export will load unchanged.
For the commercial Workflow Inspector Handoff Diagnostic and other Workflow Inspector offerings, visit workflow-inspector.com. For onboarding-specific work, use the SaaS customer onboarding product page and current product guide. Do not infer hosted capabilities from this repository's local functionality.
Keep local inputs under local-data/ and generated outputs under local-reports/; both directories are excluded by .gitignore. Use case and team reference codes, not names, email addresses, or customer notes. Reference codes are not guaranteed anonymization. Never commit customer exports, even when a validation command passes.
The validator intentionally rejects unrecognized event fields and metadata. A passing validation means the file meets this toolkit's contract, not that it is safe to publish.
python -m unittest discover -s tests -v
python scripts/check_release.py
python -m examples.python_clientThe GitHub Actions workflow runs tests and release checks with read-only repository permissions. A workflow file by itself is not evidence that a hosted CI run has passed; inspect the actual Actions result for the commit you use.
| Start here | What it covers |
|---|---|
| Quickstart | Commands, local library usage, installation, and troubleshooting. |
| Data contract | Eight CSV fields, JSON shape, validation boundaries, and normalization. |
| Metric definitions | Exact descriptive formulas, pairing rules, sample sizes, and limitations. |
| Handoff Diagnostic | What a workflow handoff diagnostic measures and how the open-source toolkit relates to the commercial Workflow Inspector Handoff Diagnostic. |
| Integration guide | Manual event mapping and the distinction between file adapters and native integrations. |
| Architecture | Local data flow and deliberately excluded capabilities. |
| Security and privacy | Reporting, sensitive-data precautions, and scope. |
| Contributing | Reproducible changes and synthetic-only examples. |
MIT applies to this repository's code and original examples. Names and branding identify the maintainer; the license is not an endorsement of forks. See NOTICE for provenance and sample-data limits. See BRAND.md for the Workflow Inspector brand and trademark boundary.