molecule-ai-plugin-ecc/tests/README.md
Molecule AI Plugin-Dev f62faec77b
Some checks failed
CI / validate (push) Failing after 2s
CI / validate (pull_request) Failing after 2s
test(ecc): add 23-test smoke suite + coverage rationale
- tests/test_ecc_smoke.py: 23 tests covering:
  - plugin.yaml schema (5 skills, 3 rules, runtimes)
  - All 3 rules/ files exist and are non-empty
  - All 5 skills/ directories + SKILL.md with valid frontmatter + body heading
  - Claude Code + deepagents adapters exist and are wired
  - known-issues.md structure (severity definitions)
  - validate-plugin.py exit 0 smoke test
- tests/README.md: explains why coverage is limited (skill+rules plugin,
  no executable hooks), and where to find integration-test guidance
2026-05-10 13:51:47 +00:00

1.5 KiB

Test Coverage Rationale — molecule-ecc

Why This Plugin Has Limited Unit-Test Coverage

molecule-ecc is a skill+rules plugin — it provides development guidelines and development skills (api-design, coding-standards, deep-research, security-review, tdd-workflow) via prose SKILL.md files and rules/*.md files.

There are no hooks, no Python business logic, and no testable adapters in this plugin. The "logic" is prose documentation.

What We Test (and Why)

What Why
plugin.yaml schema Verifies all 5 skills and 3 rules are registered
Rules files (3) Each declared rule file exists and is non-empty
Skills (5) Each skill directory + SKILL.md exists with valid YAML frontmatter and # heading
Adapters (2) Claude Code + deepagents adapters are wired
known-issues.md Severity definitions present
validate-plugin.py exit 0 Smoke test — shared CI validator passes

What We Cannot Unit-Test Here

  • SKILL.md prose content — the development guidelines are prose; their quality is a documentation review concern, not a unit-test concern.

  • Agent behavior when using skills — write integration tests in workspace-template/.

Integration Tests

If you want to test that agents actually use the ecc skills correctly, write integration tests that:

  1. Install molecule-ecc on a test workspace
  2. Ask the agent to use a specific skill (e.g., "use TDD workflow")
  3. Verify the agent follows the documented process