Skip to content

Governance and org packs

Governance packs bundle policy, ownership, and operating rhythm artifacts.

Adaptive scenario pack overlay governance

SDETKit loads adaptive scenario intelligence in deterministic layers:

  1. built-in SDETKit scenario pack,
  2. repo-local .sdetkit/adaptive/scenarios.json,
  3. additional organization or private packs listed in SDETKIT_ADAPTIVE_SCENARIO_PACKS.

Use this governance policy when an organization pack changes diagnosis behavior across multiple repositories.

Approval rules

  • New scenario codes are allowed when the pack validates against schemas/adaptive-scenario-pack.schema.json.
  • Overriding an existing scenario code requires the replacing scenario to include the explicit override-approved tag.
  • Approved overrides should include owner review notes in the PR description and a proof command that shows the new scenario ranking or proof path.
  • Security-sensitive scenarios should live in private packs and avoid embedding secret values, customer identifiers, or unredacted internal hostnames.

Validation commands

python - <<'PY'
from pathlib import Path
from sdetkit import adaptive_diagnosis

report = adaptive_diagnosis.validate_layered_scenario_packs(Path('.'))
print(report['schema_version'])
print(report['layer_count'])
print(report['overrides'])
PY

For inspection without enforcing override approval, use:

python - <<'PY'
from pathlib import Path
from sdetkit import adaptive_diagnosis

report = adaptive_diagnosis.layered_scenario_pack_report(Path('.'))
print(report['layers'])
print(report['merged_codes'][:10])
PY

Review checklist

  • Pack schema validates and loads deterministically.
  • Pack source appears in the layer metadata as repo-local or overlay-N.
  • Any duplicate scenario code is intentional and tagged with override-approved.
  • Override PR includes before/after diagnosis evidence.
  • Proof commands do not require secrets or write access.
  • Rollback is simple: remove the overlay pack or revert the scenario row.