Add 5 scenario playbooks (conflicting docs, missing schema, rejected summary, large PRs, install failure)

docs/evidence/PLAYBOOK_CONFLICTING_DOCS.md

@@ -0,0 +1,11 @@
+# Scenario Playbook: Conflicting Docs
+
+**Scenario:** Two documents give contradictory guidance.
+
+## Steps
+
+1. Identify which document is the authoritative source for the topic. Check `docs/evidence/METRICS_CURRENT.md` (metrics) or `docs/evidence/EVIDENCE_CERTIFICATE.md` (evidence).
+2. If neither is authoritative, determine which document is more recently updated or more specific.
+3. Update the less authoritative document to match the authoritative one.
+4. Add a cross-reference in both documents noting the resolution.
+5. If the conflict affects policy, log the resolution in the decision log.

docs/evidence/PLAYBOOK_INSTALL_FAILURE.md

@@ -0,0 +1,21 @@
+# Scenario Playbook: Install Failure
+
+**Scenario:** A user reports that `pip install -e ".[dev]"` fails.
+
+## Steps
+
+1. Ask for the full error message and Python version.
+2. Check if the failure is OS-specific (macOS, Linux, Windows).
+3. Common causes:
+   - Python version < 3.11
+   - Missing build dependencies (gcc, python3-dev)
+   - Conflicting packages in the user's environment
+4. Recommend a fresh virtual environment:
+   ```bash
+   python3 -m venv .venv
+   source .venv/bin/activate
+   pip install --upgrade pip
+   pip install -e ".[dev]"
+   ```
+5. If the problem persists, check `pyproject.toml` for dependency issues.
+6. Document the resolution in the issue or discussion.

docs/evidence/PLAYBOOK_LARGE_PRS.md

@@ -0,0 +1,13 @@
+# Scenario Playbook: Large PRs
+
+**Scenario:** A PR exceeds 300 lines and touches multiple areas.
+
+## Steps
+
+1. Request the author to split the PR into smaller, focused PRs.
+2. Each smaller PR should address one concern (e.g., feature change, bug fix, docs update).
+3. If splitting is not practical (e.g., a coordinated change across files):
+   - Request a design document or RFC first.
+   - Assign multiple reviewers for different areas.
+   - Set a longer review timeline.
+4. Ensure the PR description explains why the large size is necessary.

docs/evidence/PLAYBOOK_MISSING_SCHEMA.md

@@ -0,0 +1,12 @@
+# Scenario Playbook: Missing Schema Reference
+
+**Scenario:** A document references a schema file that doesn't exist or is outdated.
+
+## Steps
+
+1. Check `schemas/` for the expected schema file.
+2. If the schema exists but is outdated, update it to match the current data format.
+3. If the schema doesn't exist, create it from the data format used in the code.
+4. Update the document to reference the correct schema path.
+5. Add a test that validates the schema exists and is valid.
+6. Run `python -m openamp_foundry.cli validate` against existing data to verify.

docs/evidence/PLAYBOOK_REJECTED_SUMMARY.md

@@ -0,0 +1,14 @@
+# Scenario Playbook: Rejected Public Summary
+
+**Scenario:** A proposed public summary is rejected during review.
+
+## Steps
+
+1. Identify the specific reasons for rejection (overclaiming, missing caveats, unclear audience).
+2. Address each reason:
+   - Overclaiming: Rewrite using allowed terms from AGENTS.md.
+   - Missing caveats: Add charge-domination, no wet-lab data, simulation experimental.
+   - Unclear audience: Define who the summary is for.
+3. Re-run `python scripts/safety/check_claims.py` to verify no overclaiming.
+4. Resubmit for review.
+5. If the author disagrees with the rejection, escalate to maintainers.

tests/test_playbooks2.py

@@ -0,0 +1,21 @@
+"""Tests for playbooks batch 2."""
+from pathlib import Path
+
+FILES = [
+    "PLAYBOOK_CONFLICTING_DOCS.md",
+    "PLAYBOOK_MISSING_SCHEMA.md",
+    "PLAYBOOK_REJECTED_SUMMARY.md",
+    "PLAYBOOK_LARGE_PRS.md",
+    "PLAYBOOK_INSTALL_FAILURE.md",
+]
+
+
+def test_all_exist():
+    for f in FILES:
+        assert Path(f"docs/evidence/{f}").exists(), f"Missing: {f}"
+
+
+def test_each_has_steps():
+    for f in FILES:
+        text = Path(f"docs/evidence/{f}").read_text()
+        assert len(text) > 100