feat: add safe gate privacy checks

Adds minimal deterministic privacy and secret boundary checks to safe_pr_gate, including risky path detection, text-only marker checks, binary/deleted file skipping, and focused tests.

Author: ProfRandom92

scripts/safe_pr_gate.py

@@ -14,6 +14,15 @@
 if str(REPO_ROOT) not in sys.path:
     sys.path.insert(0, str(REPO_ROOT))
 
+RISKY_PATH_NAMES = frozenset({".env", "id_ed25519", "id_rsa"})
+RISKY_PATH_SUFFIXES = (".key", ".pem")
+PRIVATE_MARKERS = (
+    "BEGIN PRIVATE KEY",
+    "GITHUB_TOKEN=",
+    "OPENAI_API_KEY=",
+    "GEMINI_API_KEY=",
+)
+
 
 @dataclass(frozen=True, slots=True)
 class GateState:
@@ -101,6 +110,51 @@ def _path_in_prefix(path: str, prefix: str) -> bool:
     return path == normalized or path.startswith(normalized + "/")
 
 
+def _repo_relative_path(path: str) -> Path | None:
+    candidate = (REPO_ROOT / path).resolve()
+    try:
+        candidate.relative_to(REPO_ROOT)
+    except ValueError:
+        return None
+    return candidate
+
+
+def _is_risky_path(path: str) -> bool:
+    name = Path(path).name
+    return name in RISKY_PATH_NAMES or name.endswith(RISKY_PATH_SUFFIXES)
+
+
+def _read_changed_text(path: str) -> str | None:
+    candidate = _repo_relative_path(path)
+    if candidate is None or not candidate.is_file():
+        return None
+    try:
+        data = candidate.read_bytes()
+    except OSError:
+        return None
+    if b"\0" in data:
+        return None
+    try:
+        return data.decode("utf-8")
+    except UnicodeDecodeError:
+        return None
+
+
+def _privacy_problems(changed_paths: tuple[str, ...]) -> tuple[str, ...]:
+    problems: list[str] = []
+    for path in sorted(changed_paths):
+        if _is_risky_path(path):
+            problems.append(f"privacy_risky_path:{path}")
+
+        text = _read_changed_text(path)
+        if text is None:
+            continue
+        for marker in PRIVATE_MARKERS:
+            if marker in text:
+                problems.append(f"privacy_marker:{marker}:{path}")
+    return tuple(problems)
+
+
 def evaluate_gate(
     state: GateState,
     *,
@@ -125,6 +179,8 @@ def evaluate_gate(
             problems.append("changed_files_outside_allowed_prefixes")
             problems.extend(f"outside_prefix:{path}" for path in disallowed_paths)
 
+    problems.extend(_privacy_problems(state.changed_paths))
+
     return GateResult(
         ok=not problems,
         branch=state.branch,

tests/test_safe_pr_gate.py

@@ -93,6 +93,55 @@ def test_evaluate_gate_flags_paths_outside_allowed_prefixes() -> None:
     assert result.problems == ("changed_files_outside_allowed_prefixes", "outside_prefix:docs/example.md")
 
 
+def test_evaluate_gate_flags_risky_privacy_paths_in_stable_order() -> None:
+    result = evaluate_gate(
+        GateState(
+            branch="feat/safe-pr-gate",
+            status_short=(),
+            changed_paths=(
+                "secrets/id_rsa",
+                "config/.env",
+                "keys/service.key",
+                "certs/client.pem",
+            ),
+        )
+    )
+
+    assert result.ok is False
+    assert result.problems == (
+        "privacy_risky_path:certs/client.pem",
+        "privacy_risky_path:config/.env",
+        "privacy_risky_path:keys/service.key",
+        "privacy_risky_path:secrets/id_rsa",
+    )
+
+
+def test_evaluate_gate_flags_private_markers_in_changed_text_files(
+    monkeypatch: pytest.MonkeyPatch,
+    tmp_path: Path,
+) -> None:
+    text_path = tmp_path / "docs" / "example.md"
+    binary_path = tmp_path / "docs" / "binary.bin"
+    text_path.parent.mkdir()
+    text_path.write_text("GITHUB_TOKEN=example\nOPENAI_API_KEY=example\n", encoding="utf-8")
+    binary_path.write_bytes(b"\0OPENAI_API_KEY=example")
+    monkeypatch.setattr(safe_pr_gate, "REPO_ROOT", tmp_path)
+
+    result = safe_pr_gate.evaluate_gate(
+        GateState(
+            branch="feat/safe-pr-gate",
+            status_short=(),
+            changed_paths=("docs/example.md", "docs/binary.bin"),
+        )
+    )
+
+    assert result.ok is False
+    assert result.problems == (
+        "privacy_marker:GITHUB_TOKEN=:docs/example.md",
+        "privacy_marker:OPENAI_API_KEY=:docs/example.md",
+    )
+
+
 def test_parse_porcelain_paths_handles_rename_status_in_second_position() -> None:
     assert _parse_porcelain_paths(" R old-name.txt\0new-name.txt\0") == ("new-name.txt",)