Tests Governance module
Checks project policy and release gate behaviour.
Deliberate failure tests
These tests exercise mirrored AGENTS inventories, dependency restrictions, disaster-recovery freshness, library allowlists, licensing evidence and native source packaging checks. They should demonstrate that forbidden changes fail. A gate that always reports success, including on missing docs or changed source, is not protective.
Change discipline
Every new gate rule needs a valid sample and an invalid counterexample. Preserve prior historical entries and commercial evidence during upgrades. A green governance suite confirms a set of machine-checkable invariants, not that a human has read every paragraph or that a plugin update completed on the server.
Why negative cases matter
A release gate is valuable only if it rejects a source tree missing required evidence. Every new rule should have a deliberate failing fixture and prove an unchanged valid project remains allowed. Do not increase a documentation word count by repeating identical boilerplate just to satisfy a test.
Source files in this module
- tests/governance/test_agent_guides.py
- tests/governance/test_dependency_boundaries.py
- tests/governance/test_disaster_recovery_gate.py
- tests/governance/test_docs_depth_gate.py
- tests/governance/test_library_policy.py
- tests/governance/test_licensing_gate.py
- tests/governance/test_m9_m10_sources.py
Maintenance and evidence
Check all affected file references and the matching human-readable user/architecture page when source changes. Run the module’s contract or governance checks; staging, payment and native hardware tests remain independent of source inventory coverage.