Merge pull request 'Fix /media/verify apostrophe false-positive' (#60) from fix/media-verify-apostrophe-escaping into main

README.md

@@ -98,6 +98,10 @@ Integration Compose definition: `tests/compose.integration.yaml` (binds only to
 
 ## Changelog
 
+### 2.4.8 — 17.09.2026
+
+- Fixed `POST /media/verify` false-positive rejection: filenames containing a legitimate apostrophe (e.g. "La'An" in a Star Trek episode title) were blocked as "unsafe characters". Replaced the naive single-quote wrapping + apostrophe blocklist with proper `shlex.quote()` escaping for the remote ffprobe command; only newline/NUL byte injection is still rejected.
+
 ### 2.4.7 — 17.09.2026
 
 - Added `POST /media/verify`: probes a media file's real duration via `ffprobe` (running inside the existing `bazarrUHD` container on `arrapps`, which already mounts `/data` and ships ffmpeg — no new service needed) and flags it as `suspect` when it deviates from an `expected_minutes` value beyond `tolerance_pct`. Catches truncated Sonarr/Radarr imports (e.g. a re-grab that lands as 1.8 GB instead of the expected 8 GB for a UHD episode) after the file is already on the NAS.

app.py

@@ -1,7 +1,7 @@
 """Homelab Butler v2.1 – Unified API proxy for Pfannkuchen homelab.
 Reads service config from butler.yaml, credentials from Vaultwarden cache with flat-file fallback."""
 
-import os, json, asyncio, logging, time, base64, re, subprocess, ipaddress, secrets, sqlite3, math, hashlib
+import os, json, asyncio, logging, time, base64, re, subprocess, ipaddress, secrets, sqlite3, math, hashlib, shlex
 from datetime import datetime, timezone
 import httpx, yaml
 from typing import Literal
@@ -12,7 +12,7 @@ from contextlib import asynccontextmanager
 from contextvars import ContextVar
 
 log = logging.getLogger("butler")
-VERSION = "2.4.7"
+VERSION = "2.4.8"
 
 API_DIR = os.environ.get("API_KEY_DIR", "/data/api")
 VAULT_CACHE_DIR = os.environ.get("VAULT_CACHE_DIR", "/data/vault-cache")
@@ -1577,7 +1577,7 @@ async def media_verify(payload: MediaVerifyRequest, _=Depends(_verify)):
         raise HTTPException(400, f"No ffprobe container configured for host {payload.host}")
     if not re.fullmatch(r"/data/[^\x00]+\.(mkv|mp4|avi|m4v|ts)", payload.path):
         raise HTTPException(400, "Path must be an absolute /data media file")
-    if ".." in payload.path or "'" in payload.path:
+    if ".." in payload.path or "\n" in payload.path or "\x00" in payload.path:
         raise HTTPException(400, "Path contains unsafe characters")
 
     container, _default_user = MEDIA_VERIFY_PROBES[payload.host]
@@ -1590,7 +1590,7 @@ async def media_verify(payload: MediaVerifyRequest, _=Depends(_verify)):
 
     probe_cmd = (
         f"docker exec {container} ffprobe -v error "
-        f"-show_entries format=duration -of default=noprint_wrappers=1:nokey=1 '{payload.path}'"
+        f"-show_entries format=duration -of default=noprint_wrappers=1:nokey=1 {shlex.quote(payload.path)}"
     )
     rc, out, err = await asyncio.to_thread(
         _ssh, f'{target["user"]}@{target["ip"]}', f"sudo -n {probe_cmd} || {probe_cmd}", 30

tests/test_app.py

@@ -1,6 +1,7 @@
 import os
 import asyncio
 import json
+import shlex
 import time
 from datetime import datetime, timedelta, timezone
 
@@ -951,13 +952,38 @@ def test_media_verify_rejects_path_outside_data_before_ssh(monkeypatch):
     assert response.status_code == 400
 
 
+def test_media_verify_allows_apostrophe_in_filename_and_quotes_it_safely(monkeypatch):
+    monkeypatch.setattr(app, "_find_inventory_host", lambda host: {"user": "sascha", "ip": "10.2.1.100"})
+    calls = []
+
+    def fake_ssh(host, command, timeout=30):
+        calls.append((host, command, timeout))
+        return 0, "2967.0\n", ""
+
+    monkeypatch.setattr(app, "_ssh", fake_ssh)
+    path = "/data/FHD/serien/Star Trek - Strange New Worlds (2022)/Season 04/Once La'An a Time.mkv"
+    with TestClient(app.app) as client:
+        response = client.post(
+            "/media/verify",
+            headers={"Authorization": "Bearer test-token"},
+            json={"host": "arrapps", "path": path, "expected_minutes": 49.5},
+        )
+    assert response.status_code == 200
+    # shlex.quote must produce a command the remote shell parses as ONE argument,
+    # i.e. no unescaped apostrophe breaks out of quoting.
+    executed_cmd = calls[0][1]
+    quoted = shlex.quote(path)
+    assert quoted in executed_cmd
+    assert shlex.split(executed_cmd)[-1] == path
+
+
 def test_media_verify_rejects_shell_metacharacters_before_ssh(monkeypatch):
     monkeypatch.setattr(app, "_ssh", lambda *_args, **_kwargs: (_ for _ in ()).throw(AssertionError("must not SSH")))
     with TestClient(app.app) as client:
         response = client.post(
             "/media/verify",
             headers={"Authorization": "Bearer test-token"},
-            json={"host": "arrapps", "path": "/data/UHD/serien/a'; rm -rf /'.mkv"},
+            json={"host": "arrapps", "path": "/data/UHD/serien/a\x00.mkv"},
         )
     assert response.status_code == 400