Add 10 infrastructure files

docs/evidence/CHECKLIST_DEPENDENCY_UPDATE.md

@@ -0,0 +1,16 @@
+# Dependency Update Checklist
+
+Steps for updating a dependency.
+
+- [ ] Check the new version's changelog for breaking changes.
+- [ ] Update the version in pyproject.toml.
+- [ ] Run `pip install -e ".[dev]"` to install the update.
+- [ ] Run `python3 -m pytest -q` to verify nothing is broken.
+- [ ] If tests fail, fix compatibility issues.
+- [ ] Update any documentation that references the old version.
+- [ ] If the dependency has a security fix, note it in the PR.
+
+## Frequency
+- Security updates: apply within 1 week.
+- Minor updates: apply within 1 month.
+- Major updates: evaluate on a case-by-case basis.

docs/evidence/GUIDE_PRIVACY_THREAT_MODEL.md

@@ -0,0 +1,16 @@
+# Artifact Privacy Threat Model
+
+Potential privacy threats related to pipeline artifacts.
+
+| Threat | Risk | Mitigation |
+|--------|:----:|------------|
+| Candidate sequences in logs | Low | Logs are local-only |
+| Lab results contain PII | Medium | Lab partner agreement, data return schema excludes PII |
+| Evidence certificates shared without context | Low | All certs include disclaimer |
+| Manifest reveals internal paths | Low | Paths are relative, no user info |
+
+## Principles
+- Pipeline outputs should not contain PII.
+- If PII is present, it must be clearly labeled and handled per data governance policy.
+- All artifacts include a disclaimer stating they are computational predictions.
+- Artifacts shared externally should be reviewed for sensitive content.

docs/evidence/GUIDE_STRUCTURED_LOGGING_OPTION.md

@@ -0,0 +1,23 @@
+# Structured Logging Option
+
+How to enable structured logging.
+
+## Current State
+Pipeline uses print statements and subprocess output. There is no
+structured logging framework.
+
+## Recommendation
+Add Python's `logging` module with structured formatting:
+```python
+import logging
+logging.basicConfig(
+    format="%(asctime)s  %(levelname)s  %(name)s:%(lineno)d  %(message)s",
+    level=logging.INFO,
+)
+```
+
+## Benefits
+- Consistent log format across all modules.
+- Log levels (DEBUG, INFO, WARNING, ERROR) enable filtering.
+- Structured format enables log parsing and analysis.
+- No external dependencies required.

docs/evidence/POLICY_DEPENDENCY_RISK.md

@@ -0,0 +1,21 @@
+# Dependency Risk Review Policy
+
+How to review dependency risks.
+
+## Risk Levels
+| Level | Criteria | Review Required |
+|:-----:|----------|:---------------:|
+| Low | Well-known library, stable API, permissive license | None |
+| Medium | Less common library, breaking changes possible | Maintainer |
+| High | New dependency, restrictive license, security-sensitive | Project lead |
+
+## Required Checks
+- License compatibility with Apache 2.0
+- Known security vulnerabilities
+- Maintenance status (recent releases, issue response)
+- Number of transitive dependencies
+
+## Prohibited
+- Dependencies with GPL or AGPL licenses (unless absolutely necessary and reviewed)
+- Dependencies with known critical vulnerabilities
+- Unmaintained dependencies (no release in > 2 years)

docs/evidence/POLICY_HASH_ALGORITHM.md

@@ -0,0 +1,20 @@
+# Hash Algorithm Policy
+
+Hash algorithms used in the pipeline.
+
+## Current Algorithms
+| Algorithm | Used For | Strength |
+|-----------|----------|:--------:|
+| SHA-256 | Evidence certificates, manifests, chain of custody | Strong |
+| SHA-256 | Input file hashing for reproducibility | Strong |
+
+## Requirements
+- All hashes must use SHA-256 or stronger.
+- MD5 and SHA-1 are forbidden for security-sensitive uses.
+- Hash algorithms must be documented in the relevant schema.
+- Hash values must be hex-encoded, lowercase.
+
+## Verification
+- Hash verification is built into chain-of-custody scripts.
+- Run `scripts/verify_lab_peptides.py` to verify sequence hashes.
+- Manifests include hashes for all input files.

docs/evidence/POLICY_SAFE_ARCHIVE.md

@@ -0,0 +1,16 @@
+# Safe Archive Extraction Policy
+
+How to safely extract archives in the pipeline.
+
+## Rules
+- Only extract archives from trusted sources.
+- Validate archive contents before extraction.
+- Prevent path traversal attacks (no `../` in extracted paths).
+- Set a maximum file size and count.
+- Extract to a dedicated directory, not to an existing directory tree.
+- Verify extracted file hashes against expected values when available.
+
+## Implementation
+- The `build_lab_batch_pack.py` script creates ZIP archives.
+- Extraction should use `zipfile.ZipFile.extractall()` with path validation.
+- Path traversal is prevented by checking that all extracted paths are within the target directory.

docs/evidence/PROPOSAL_CLI_COMPLETION.md

@@ -0,0 +1,25 @@
+# CLI Completion Support Proposal
+
+Adding shell completion support for the CLI.
+
+## Motivation
+Shell completion improves developer experience by showing available
+commands, flags, and values.
+
+## Approach
+Use Python's `argparse` with `argcomplete` for shell completion.
+
+## Required Changes
+1. Add `argcomplete` as an optional dependency.
+2. Add shell completion entry point in `pyproject.toml`.
+3. Register argparse parsers with `argcomplete.autocomplete()`.
+4. Document how to enable completion in the shell.
+
+## Completion Types
+- Command names: `rank`, `validate`, `bench`, etc.
+- Flag names: `--candidates`, `--out`, `--config`, etc.
+- File paths: for `--candidates`, `--out` flags.
+- Enum values: for `--ranking-mode` (ensemble, expert).
+
+## Status
+Proposed — not yet implemented.

tests/test_manifest_completeness.py

@@ -0,0 +1,23 @@
+"""Tests for archive manifest completeness."""
+import json
+from pathlib import Path
+
+
+def test_run_manifest_has_required_keys():
+    """Verify run_manifest.schema.json has required fields."""
+    schema = json.loads(Path("schemas/run_manifest.schema.json").read_text())
+    required = {"run_id", "pipeline_version", "config_hash", "generated_at", "inputs", "outputs"}
+    if "required" in schema:
+        assert required.issubset(set(schema["required"])), "Missing required fields in schema"
+
+
+def test_candidate_schema_has_required_keys():
+    schema = json.loads(Path("schemas/candidate.schema.json").read_text())
+    assert "candidate_id" in str(schema)
+    assert "scores" in str(schema)
+
+
+def test_lab_result_schema_has_required_keys():
+    schema = json.loads(Path("schemas/lab_result.schema.json").read_text())
+    for field in ["candidate_id", "assay_type", "result_value", "result_unit"]:
+        assert field in str(schema), f"Missing {field} from lab_result schema"

tests/test_path_traversal.py

@@ -0,0 +1,20 @@
+"""Tests for path traversal in archive builders."""
+import zipfile
+import tempfile
+from pathlib import Path
+
+
+def test_zip_path_traversal_detected():
+    """Verify that path traversal via ../ in zip entries can be detected."""
+    with tempfile.TemporaryDirectory() as tmp:
+        malicious_zip = Path(tmp) / "malicious.zip"
+        with zipfile.ZipFile(malicious_zip, "w") as zf:
+            zf.writestr("../../etc/passwd", "malicious")
+        detected = False
+        with zipfile.ZipFile(malicious_zip, "r") as zf:
+            for name in zf.namelist():
+                resolved = (Path(tmp) / name).resolve()
+                if not str(resolved).startswith(str(Path(tmp).resolve())):
+                    detected = True
+        assert detected, "Path traversal was not detected"
+        assert "../../" in name, "Path traversal pattern should contain ../"

tests/test_policies_batch.py

@@ -0,0 +1,17 @@
+"""Tests for policies and infrastructure batch."""
+from pathlib import Path
+
+FILES = [
+    "CHECKLIST_DEPENDENCY_UPDATE.md",
+    "GUIDE_PRIVACY_THREAT_MODEL.md",
+    "POLICY_DEPENDENCY_RISK.md",
+    "POLICY_HASH_ALGORITHM.md",
+    "POLICY_SAFE_ARCHIVE.md",
+    "GUIDE_STRUCTURED_LOGGING_OPTION.md",
+    "PROPOSAL_CLI_COMPLETION.md",
+]
+
+
+def test_all_exist():
+    for f in FILES:
+        assert Path("docs/evidence", f).exists(), f"Missing: {f}"

tests/test_stdout_stderr.py

@@ -0,0 +1,48 @@
+"""Tests that CLI commands properly separate stdout and stderr."""
+import subprocess
+import sys
+
+import pytest
+
+
+def _run_cmd(*args: str) -> subprocess.CompletedProcess:
+    return subprocess.run(
+        [sys.executable, "-m", "openamp_foundry.cli", *args],
+        capture_output=True, text=True,
+        env={"PYTHONPATH": "src"},
+    )
+
+
+def test_rank_output_goes_to_stdout():
+    r = _run_cmd("rank", "--help")
+    assert r.returncode == 0
+    assert len(r.stdout) > 0
+
+
+def test_validate_output_goes_to_stdout():
+    r = _run_cmd("validate", "--help")
+    assert r.returncode == 0
+    assert len(r.stdout) > 0
+
+
+def test_bench_help_goes_to_stdout():
+    r = _run_cmd("bench", "--help")
+    assert r.returncode == 0
+    assert len(r.stdout) > 0
+
+
+def test_missing_file_error_goes_to_stderr():
+    r = _run_cmd("rank", "--candidates", "/nonexistent.csv", "--out", "/tmp/x.jsonl")
+    assert r.returncode != 0
+    assert len(r.stderr) > 0 or "Error" in r.stdout
+
+
+@pytest.mark.parametrize("cmd", [
+    ["rank", "--help"],
+    ["validate", "--help"],
+    ["bench", "--help"],
+])
+def test_help_uses_stdout(cmd):
+    r = _run_cmd(*cmd)
+    assert r.returncode == 0
+    assert len(r.stdout) > 0