test(presets): probe CLI/docs parity for all four presets

[0xjgv]

Summary

• Probes found real drift: progressive preset was missing from test assertions and README docs
• Tightened rejection-message check from baseline|strict|legacy to include progressive
• Added 3 new tests: descriptions × 4, defaults key-values × 21, progressive blocking-gates integration
• Added BDD scenario for cli-presets-parity in interlock_cli.feature
• Fixed README: added progressive to preset comment and bullet list

Files changed

• tests/test_cli.py — 3 new parametrized/integration tests, 2 tightened assertions, Preset import
• tests/features/interlock_cli.feature — new cli-presets-parity scenario
• README.md — progressive preset in comment and description list

Test plan

• uv run pytest -q tests/test_cli.py tests/step_defs/test_interlock_cli.py — 69 passed
• Pre-commit (ruff fix, format, basedpyright) — all green

[gemini-code-assist]

Code Review

This pull request introduces a new 'progressive' preset, which acts as an autopilot ratchet by using blocking gates while automatically advancing thresholds via a baseline file. The changes include documentation updates, new Gherkin scenarios for CLI parity, and expanded unit tests for preset configurations. Feedback focuses on increasing test coverage for the 'strict' and 'progressive' presets by verifying additional configuration keys such as 'enforce_behavior_attribution' and 'run_acceptance_in_check'.

[gemini-code-assist]

The strict preset explicitly enables enforce_behavior_attribution. It should be included in this parity test to ensure the preset is correctly configured.

Suggested change

	# strict: all blocking gates on, mutation incremental
	("strict", "enforce_crap", True),
	("strict", "enforce_mutation", True),
	("strict", "run_mutation_in_ci", True),
	("strict", "mutation_ci_mode", "incremental"),
	("strict", "run_acceptance_in_check", True),
	("strict", "require_acceptance", True),
	("strict", "coverage_min", 90),
	# strict: all blocking gates on, mutation incremental
	("strict", "enforce_crap", True),
	("strict", "enforce_behavior_attribution", True),
	("strict", "enforce_mutation", True),
	("strict", "run_mutation_in_ci", True),
	("strict", "mutation_ci_mode", "incremental"),
	("strict", "run_acceptance_in_check", True),
	("strict", "require_acceptance", True),
	("strict", "coverage_min", 90),

[gemini-code-assist]

The progressive preset also explicitly enables enforce_behavior_attribution. Including it here ensures full coverage of the preset's intended blocking behavior.

        # progressive: blocking gates on, permissive floors (ratcheted at runtime)
        ("progressive", "enforce_crap", True),
        ("progressive", "enforce_behavior_attribution", True),
        ("progressive", "enforce_mutation", True),
        ("progressive", "run_mutation_in_ci", True),
        ("progressive", "mutation_ci_mode", "incremental"),
        ("progressive", "run_acceptance_in_check", True),
        ("progressive", "require_acceptance", True),
        ("progressive", "coverage_min", 0),  # floor; ratcheted by baseline.json

[gemini-code-assist]

This integration test verifies that the progressive preset correctly wires blocking gates. It should also verify enforce_behavior_attribution and run_acceptance_in_check, which are key parts of the preset's configuration.

    assert re.search(r"^\s*preset\s+progressive\s*$", out, re.MULTILINE), out
    assert re.search(r"^\s*enforce_crap\s+True \(preset-derived\)\s*$", out, re.MULTILINE), out
    assert re.search(r"^\s*enforce_behavior_attribution\s+True \(preset-derived\)\s*$", out, re.MULTILINE), out
    assert re.search(r"^\s*enforce_mutation\s+True \(preset-derived\)\s*$", out, re.MULTILINE), out
    assert re.search(r"^\s*run_mutation_in_ci\s+True \(preset-derived\)\s*$", out, re.MULTILINE), out
    assert re.search(r"^\s*run_acceptance_in_check\s+True \(preset-derived\)\s*$", out, re.MULTILINE), out
    assert re.search(r"^\s*require_acceptance\s+True \(preset-derived\)\s*$", out, re.MULTILINE), out