ci: drop pytest pipeline, keep ruff/bandit/docs

The clean release ships without a tests/ folder, so the original CI's
pytest+coverage steps had nothing to run. Strip them from ci.yml, drop
pytest extras and ini_options from pyproject, and remove dead test
references from CONTRIBUTING.md and README.md.

CI now runs ruff, bandit, and a docs-presence check across Python
3.10/3.11/3.12; inspect-smoke on PRs is unchanged.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Author: n-papaioannou

.github/workflows/ci.yml

@@ -6,7 +6,7 @@ on:
     branches: [main, "fix/**"]
 
 jobs:
-  test:
+  checks:
     runs-on: ubuntu-latest
     strategy:
       matrix:
@@ -24,13 +24,7 @@ jobs:
         run: python -m pip install -e ".[dev]"
 
       - name: Lint
-        run: ruff check ifixai tests
-
-      - name: Test with coverage
-        run: pytest tests/ ifixai/tests/ --cov=ifixai --cov-report=term-missing --cov-fail-under=60 -q --tb=short -m "not integration"
-
-      - name: Integration tests (disk/fixture-dependent)
-        run: pytest tests/ -q --tb=short -m integration
+        run: ruff check ifixai
 
       - name: Security scan
         run: bandit -r ifixai -ll
@@ -44,7 +38,7 @@ jobs:
 
   inspect-smoke:
     runs-on: ubuntu-latest
-    needs: test
+    needs: checks
     if: ${{ github.event_name == 'pull_request' }}
     steps:
       - uses: actions/checkout@v4

CONTRIBUTING.md

@@ -1,6 +1,6 @@
 # Contributing to ifixai
 
-Thanks for considering a contribution. This guide covers the mechanics of adding inspections, fixtures, providers, and running the test suite. For the behavioural contract (what the project expects from the code you write), see the root `CLAUDE.md` if it exists plus `.claude/rules/common/*.md` in the repository.
+Thanks for considering a contribution. This guide covers the mechanics of adding inspections, fixtures, and providers. For the behavioural contract (what the project expects from the code you write), see the root `CLAUDE.md` if it exists plus `.claude/rules/common/*.md` in the repository.
 
 ## Environment setup
 
@@ -18,13 +18,10 @@ The `pre-commit install` step wires up local hooks (`gitleaks`, `ruff`, a `.env`
 Verify:
 
 ```bash
-ruff check ifixai tests
-python -m pytest --cov=ifixai --cov-report=term
+ruff check ifixai
 ```
 
-The coverage floor is declared in `.github/workflows/ci.yml`. The policy is simple: the floor is pinned at `floor(current_baseline) - 2` percentage points so a flaky test cannot accidentally erode coverage, and is ratcheted upward when a sustained improvement lands. Drop-below fails CI.
-
-## Adding a test (inspection)
+## Adding an inspection
 
 Each inspection is one file under `ifixai/tests/bNN_short_name.py`. The minimum contract:
 
@@ -35,14 +32,6 @@ Each inspection is one file under `ifixai/tests/bNN_short_name.py`. The minimum
 5. Register the inspection in `ifixai/tests/registry.py` (import + add to `ALL_SPECS` + `create_inspection` switch).
 6. Update `ifixai/scoring/category_weights.py` only if the inspection belongs to the strategic set.
 
-### Inspection testing
-
-Every inspection should have a companion test under `tests/test_bNN_*.py` covering at least:
-
-- `test_{inspection}_spec_invariants_unchanged` — asserts immutable fields (id, category, threshold, weight, strategic-ness).
-- Happy-path evidence generation against a fixture.
-- Failure path (provider error, empty response, malformed output).
-
 ## Authoring a fixture
 
 Fixtures are YAML files under `ifixai/fixtures/`. Validate against `ifixai/fixtures/schema.json`. A fixture MUST supply every key listed in the union of every registered inspection's `required_fixture_keys`.
@@ -59,22 +48,12 @@ Providers implement the `ChatProvider` protocol from `ifixai/providers/base.py`.
 2. Register the provider string in `ifixai/providers/resolver.py`.
 3. Add an optional dependency extra in `pyproject.toml` under `[project.optional-dependencies]` so users install only what they need.
 4. Do not swallow exceptions silently. If the provider has idiomatic error types, translate them into `ProviderError` (or a subclass).
-5. Add `tests/test_<your_provider>.py` with stub / mock coverage.
 
-## Running tests locally
+## Local checks
 
 ```bash
-# full suite
-python -m pytest
-
-# one module
-python -m pytest tests/test_your_thing.py
-
-# with coverage
-python -m pytest --cov=ifixai --cov-report=term-missing
-
-# ruff
-ruff check ifixai tests
+# lint
+ruff check ifixai
 
 # type check
 mypy ifixai
@@ -91,19 +70,17 @@ Follow the Conventional Commits-style prefixes used by this project:
 - `fix:` bug fix
 - `refactor:` behaviour-preserving change
 - `docs:` documentation only
-- `test:` test-only changes
 - `chore:` tooling / housekeeping
 - `perf:` performance improvement
 - `ci:` CI configuration
 
-Keep commits small and atomic. Include a test with every `fix:` or `feat:` that changes observable behaviour.
+Keep commits small and atomic.
 
 ## Pull requests
 
 - Target branch: `main`.
-- Include a test plan in the PR body.
-- Confirm `ruff`, `pytest`, and `bandit` all pass locally before requesting review. `mypy` is advisory (run it locally if you touched typed surfaces, but it is not a CI gate).
-- For any inspection / fixture / provider change, paste one worked example scorecard snippet (JSON or Markdown) into the PR body. The `ifixai-results/` directory is gitignored and cannot be updated as part of a PR.
+- Confirm `ruff` and `bandit` pass locally before requesting review. `mypy` is advisory (run it locally if you touched typed surfaces, but it is not a CI gate).
+- For any inspection / fixture / provider change, paste one worked example scorecard snippet (JSON or Markdown) into the PR body.
 
 ## Where to ask
 

README.md

@@ -184,13 +184,10 @@ healthcare, software engineering, customer support, finance, legal,
 government, anything. Industry knowledge lives **only in user authored
 fixture YAML**, never in test code.
 
-A guardrail test ([`tests/test_industry_agnosticism.py`](tests/test_industry_agnosticism.py))
-scans every inspection, rule, judge reference, and analytic rubric against a
-word and phrase level industry blocklist on every CI run. New contributors
-who try to hardcode healthcare, legal, or finance prompts are blocked at PR
-time. The blocklist is literal substring matching; a deliberate contributor
-can still route domain content through fixture YAML (the intended path)
-rather than hardcoded strings.
+Inspection prompts, rules, judge references, and analytic rubrics are kept
+free of domain specific content by convention. Industry knowledge belongs in
+user authored fixture YAML — keeping it out of the inspection code is what
+lets the same suite run meaningfully against agents in any domain.
 
 Three illustrative example fixtures live under
 [`ifixai/fixtures/examples/`](ifixai/fixtures/examples/) demonstrating the
@@ -371,8 +368,7 @@ Custom providers can be passed directly by implementing `ChatProvider` from [ifi
 
 ```bash
 pip install -e ".[dev]"
-pytest tests/
-ruff check ifixai tests
+ruff check ifixai
 ```
 
 ## License

pyproject.toml

@@ -50,9 +50,6 @@ all = [
     "huggingface-hub>=0.20",
 ]
 dev = [
-    "pytest>=7.0",
-    "pytest-asyncio>=0.23",
-    "pytest-cov>=4.0",
     "ruff>=0.4",
     "mypy>=1.0",
     "bandit>=1.7",
@@ -69,11 +66,6 @@ Documentation = "https://github.com/ifixai-ai/diagnostic#readme"
 Repository = "https://github.com/ifixai-ai/diagnostic"
 Issues = "https://github.com/ifixai-ai/diagnostic/issues"
 
-[tool.pytest.ini_options]
-markers = [
-    "integration: tests that depend on on-disk fixtures, shipped YAML, or recorded cassettes",
-]
-
 [tool.setuptools.packages.find]
 include = ["ifixai*"]