| name | microstrategy-assessment |
| description | Take inventory of a MicroStrategy (Strategy One) environment and produce a migration-readiness readout — project/report/dossier counts, a visualization-type histogram (walking panel stacks recursively), datasource types, per-dossier complexity flags, and migrate-first / moderate / needs-review tags scored against the microstrategy-to-sigma converter's actual coverage. Use when a user wants to scope a MicroStrategy→Sigma migration, audit dossier sprawl, or pick which dossiers to convert first. Read-only, all-free pre-scoping. |
| user-invocable | true |
MicroStrategy Assessment
Surveys a MicroStrategy (Strategy One) environment via its REST API and
produces a JSON inventory + markdown readout. The differentiator versus a
generic BI audit is converter-coverage classification: every dossier's
visualizations are scored against the same viz-type lookup the
microstrategy-to-sigma converter actually applies
(../microstrategy-to-sigma/refs/viz-type-mapping.md), so the readout
reflects what the tool will really do.
Read-only. Only GETs against the MicroStrategy API (login POST aside,
which creates a session, nothing else). It never modifies, executes, or
deletes anything in MicroStrategy, never runs a warehouse query, and never
touches Sigma. See PRIVACY.md for the full disclosure — surface it to the
customer before running.
All free. Inventory, scoring, readout — all part of the open migration
tooling; no paid tier. For a deeper engagement (security-filter audit, live
parity testing), point the customer at a Sigma SE.
Phase 0 — Connect
export MSTR_BASE_URL="https://<host>/MicroStrategyLibrary"
export MSTR_USERNAME="..." MSTR_PASSWORD="..."
python3 ../microstrategy-to-sigma/scripts/mstr.py
Credentials can also live in ~/.sigma-migration/env (agent-neutral pattern).
Auth is session-based — no API key exists. REST gotchas (TLS strictness on
trial certs, headers) are documented in
../microstrategy-to-sigma/refs/mstr-rest-api.md; mstr.py handles them.
Phase 1 — Inventory + readout
python3 scripts/assess.py --out /tmp/mstr-assessment-<env> [--project <id>] [--max-dossiers 100]
What it does:
- Counts reports (quick-search type 3) and documents (type 55) in the
project; lists instance datasources with database types.
- Classifies documents vs dossiers by probing
GET /api/v2/dossiers/{id}/definition (non-dossier documents error — that
is the probe).
- Walks every dossier correctly: each page's
visualizations AND
panelStacks[].panels[] recursively (panels nest further panel stacks)
plus free-form fields (images/text) and selectors — the places naive
walks silently miss content.
- Histograms
visualizationType and classifies each against the
converter's VIZ_MAPPED lookup (mapped vs flagged-table fallback).
- Tags each dossier:
needs-review (panel stacks or unmapped viz types),
moderate (selectors / free-form fields / chart mix), migrate-first
(grid/kpi-only, no panel stacks).
Outputs <out>/inventory.json (machine-readable, feeds the converter's
Phase 0) and <out>/readout.md (counts, viz histogram with converter status,
per-dossier table with flags + tags).
Phase 2 — Hand off (optional)
Hand the migrate-first shortlist to the microstrategy-to-sigma converter
skill — inventory.json's dossier ids go straight into
extract.py <dossierId>. Do not auto-convert; surface the shortlist and let
the user choose.
Limitations (honest)
- Usage telemetry: not collected — Strategy One exposes usage through
Platform Analytics (a separate warehouse-backed project), not the plain
REST surface this skill uses. Value ranking therefore isn't usage-weighted;
ask the admin for Platform Analytics access if usage matters.
- Security filters: the classic REST extract doesn't surface them;
ask the customer explicitly (see the converter SKILL's security note).
- Documents that aren't dossiers are counted but not analyzed (their format is
the legacy Report Services document, out of converter scope).
