Merge branch 'security-hardening' into main

Author: crispasr integration

README.md

@@ -642,6 +642,7 @@ def initialize_session(request: gr.Request):
             "has_sandbox_access": getattr(user, "has_sandbox_access", False),
             "umk": umk,
             "max_model_size_b": _max_model_b,
+            "must_change_password": getattr(user, "must_change_password", False),
         }
 
         ensure_user_storage_dirs(username)
@@ -720,10 +721,16 @@ def initialize_session(request: gr.Request):
         ],
     )
 
-    # Logout: redirect to Gradio's built-in /logout endpoint
+    # Logout: clear server-side UMK cache, then redirect to Gradio /logout
+    def _on_logout(user_state):
+        username = (user_state or {}).get("username")
+        if username:
+            from db_ops import clear_cached_umk
+            clear_cached_umk(username)
+
     logout_btn.click(
-        fn=None,
-        inputs=None,
+        fn=_on_logout,
+        inputs=[session_state],
         outputs=None,
         js="() => { window.location.href = '/logout'; }",
     )

app.py

@@ -356,6 +356,13 @@ def _handler(args: dict):
         if not url.startswith(("http://", "https://")):
             return ToolResult(text=f"Invalid URL: {url}", error=True)
 
+        # SSRF guard: block private/loopback/metadata IPs
+        from ssrf_guard import is_safe_url
+
+        ok, reason = is_safe_url(url)
+        if not ok:
+            return ToolResult(text=f"URL blocked ({reason}): {url}", error=True)
+
         print(f"[TOOLS] 🔬 analyze_image_url | url={url!r}")
 
         # Download image to temp file
@@ -505,8 +512,7 @@ def no_creds_error(*args, **kwargs):  # type: ignore[no-untyped-def]
             "exchange_contact_edit": no_creds_error,
         }
 
-    _masked = {k: ("***" if k == "password" else v) for k, v in creds.items()}
-    logger.info("[EXCH] Credentials loaded for user_id=%s: %s", user_id, _masked)
+    logger.info("[EXCH] Credentials loaded for user_id=%s: email=%s", user_id, creds.get("email", "?"))
 
     def handle_exchange_list(args: dict) -> str:
         logger.debug("[EXCH] exchange_list called with args=%s", args)
@@ -527,10 +533,22 @@ def handle_exchange_send(args: dict) -> str:
             "[EXCH] exchange_send called with args=%s",
             {k: ("***" if k == "body" else v) for k, v in args.items()},
         )
-        # schema uses 'mailbox' for sender override; function expects 'email'
+        # Security: force draft=True unless the user explicitly set draft=False.
+        # This prevents prompt-injection attacks from silently sending emails.
         send_args = dict(args)
         if "mailbox" in send_args:
             send_args["email"] = send_args.pop("mailbox")
+        if not send_args.get("draft") is False:
+            # Default to saving as draft for safety
+            send_args["draft"] = True
+            result = exchange_cli.send_mail(credentials_dict=creds, **send_args)
+            logger.info("[EXCH] exchange_send (draft) result: %s", result)
+            return (
+                f"{result}\n\n"
+                "Die E-Mail wurde als ENTWURF gespeichert (Sicherheitsmassnahme). "
+                "Der Benutzer muss den Entwurf manuell in Outlook senden oder "
+                "die Anfrage mit draft=false wiederholen."
+            )
         result = exchange_cli.send_mail(credentials_dict=creds, **send_args)
         logger.info("[EXCH] exchange_send result: %s", result)
         return result
@@ -659,7 +677,15 @@ def run_chat(
                     "brave_answers", "duckduckgo", "searxng").
     inject_date: prepend today's date to the system prompt (default True).
     """
-    # VALIDATION
+    # VALIDATION — enforce password rotation before any chat usage
+    if (user_state or {}).get("must_change_password"):
+        yield (
+            "🔒 **Passwortänderung erforderlich**\n\n"
+            "Ihr Passwort muss geändert werden, bevor Sie den Chat nutzen können.\n"
+            "Bitte gehen Sie zum Tab **Historie → Passwort ändern**."
+        )
+        return
+
     allowed, reason = is_provider_allowed(provider, user_state)
     if not allowed:
         yield f"⛔ **Zugriff verweigert**\n\n{reason}\n\nBitte wählen Sie einen EU-Provider (Scaleway, Mistral, Nebius)."
@@ -1246,7 +1272,8 @@ def run_chat(
             return
 
         logger.exception(f"Chat error with {provider}: {e!s}")
-        yield f"🔥 Fehler ({provider}): {e!s}"
+        # Sanitize: don't leak internal paths or stack traces to the user
+        yield f"🔥 Fehler bei der Kommunikation mit {provider}. Bitte erneut versuchen."
 
 
 # ==========================================
@@ -1505,7 +1532,7 @@ def prepare_api_payload(content, role, is_active_user_msg=False):
         import traceback
 
         traceback.print_exc()
-        hist[-1]["content"] = f"🔥 System-Fehler: {e!s}"
+        hist[-1]["content"] = "🔥 System-Fehler. Bitte erneut versuchen."
         yield hist
 
 

chat_handlers.py

@@ -73,22 +73,22 @@ def check_storage_health(path, timeout=3):
 # API KEYS
 # ==========================================
 API_KEYS = {
-    "GLADIA": os.environ.get("GLADIA_API_KEY", "your_key"),
-    "SCALEWAY": os.environ.get("SCALEWAY_API_KEY", "your_key"),
-    "NEBIUS": os.environ.get("NEBIUS_API_KEY", "your_key"),
-    "MISTRAL": os.environ.get("MISTRAL_API_KEY", "your_key"),
-    "OPENROUTER": os.environ.get("OPENROUTER_API_KEY", "your_key"),
-    "GROQ": os.environ.get("GROQ_API_KEY", "your_key"),
-    "POE": os.environ.get("POE_API_KEY", "your_poe_key_here"),
-    "DEEPGRAM": os.environ.get("DEEPGRAM_API_KEY", "your_key"),
-    "ASSEMBLYAI": os.environ.get("ASSEMBLYAI_API_KEY", "your_key"),
-    "OPENAI": os.environ.get("OPENAI_API_KEY", "your_key"),
-    "COHERE": os.environ.get("COHERE_API_KEY", "your_key"),
-    "TOGETHER": os.environ.get("TOGETHER_API_KEY", "your_key"),
-    "OVH": os.environ.get("OVH_API_KEY", "your_key"),
-    "CEREBRAS": os.environ.get("CEREBRAS_API_KEY", "your_key"),
-    "GOOGLEAI": os.environ.get("GOOGLEAI_API_KEY", "your_key"),
-    "ANTHROPIC": os.environ.get("ANTHROPIC_API_KEY", "your_key"),
+    "GLADIA": os.environ.get("GLADIA_API_KEY", ""),
+    "SCALEWAY": os.environ.get("SCALEWAY_API_KEY", ""),
+    "NEBIUS": os.environ.get("NEBIUS_API_KEY", ""),
+    "MISTRAL": os.environ.get("MISTRAL_API_KEY", ""),
+    "OPENROUTER": os.environ.get("OPENROUTER_API_KEY", ""),
+    "GROQ": os.environ.get("GROQ_API_KEY", ""),
+    "POE": os.environ.get("POE_API_KEY", ""),
+    "DEEPGRAM": os.environ.get("DEEPGRAM_API_KEY", ""),
+    "ASSEMBLYAI": os.environ.get("ASSEMBLYAI_API_KEY", ""),
+    "OPENAI": os.environ.get("OPENAI_API_KEY", ""),
+    "COHERE": os.environ.get("COHERE_API_KEY", ""),
+    "TOGETHER": os.environ.get("TOGETHER_API_KEY", ""),
+    "OVH": os.environ.get("OVH_API_KEY", ""),
+    "CEREBRAS": os.environ.get("CEREBRAS_API_KEY", ""),
+    "GOOGLEAI": os.environ.get("GOOGLEAI_API_KEY", ""),
+    "ANTHROPIC": os.environ.get("ANTHROPIC_API_KEY", ""),
     "BFL": os.environ.get("BFL_API_KEY", ""),
     "REQUESTY": os.environ.get("REQUESTY_API_KEY", ""),
     "LANGDOCK": os.environ.get("LANGDOCK_API_KEY", ""),

config.py

@@ -250,7 +250,7 @@ def run_embed_text(texts: list[str], model: str = "all-MiniLM-L6-v2") -> list[li
 
     import json
 
-    cmd = [binary, "-m", model, "--json", "--auto-download"] + list(texts)
+    cmd = [binary, "-m", model, "--json", "--auto-download", "--"] + list(texts)
 
     try:
         proc = subprocess.run(  # nosec B603

crispembed_tool.py

@@ -138,6 +138,12 @@ def get_cached_umk(username: str):
         return _UMK_CACHE.get(username)
 
 
+def clear_cached_umk(username: str):
+    """Remove the cached UMK for *username* on logout."""
+    with _UMK_LOCK:
+        _UMK_CACHE.pop(username, None)
+
+
 def authenticate_user(username, password):
     """
     Authenticate user AND create keychain if needed.
@@ -499,6 +505,9 @@ def change_password_secure(user_id, old_password, new_password):
 
         # Standard password update
         user.password_hash = hash_password(new_password)
+        # Clear the must_change_password flag after successful rotation
+        if hasattr(user, "must_change_password"):
+            user.must_change_password = False
         db.commit()
         return True, "✅ Passwort geändert & Key neu verschlüsselt (Daten bleiben sicher)"
     except Exception as e:
@@ -629,14 +638,16 @@ def get_decrypted_vision(vision_id: int, user_id: int, user_state=None):
         if vis.is_encrypted:
             try:
                 # Decrypt with user key, fallback to global
-                vis.result = crypto.decrypt_text(str(vis.result), key=umk)  # type: ignore[assignment]
+                orig_result = str(vis.result)
+                vis.result = crypto.decrypt_text(orig_result, key=umk)  # type: ignore[assignment]
                 if vis.result == "[Decryption Failed]" and umk != crypto.global_key:
-                    vis.result = crypto.decrypt_text(str(vis.result), key=crypto.global_key)  # type: ignore[assignment]
+                    vis.result = crypto.decrypt_text(orig_result, key=crypto.global_key)  # type: ignore[assignment]
 
                 if vis.prompt:
-                    vis.prompt = crypto.decrypt_text(str(vis.prompt), key=umk)  # type: ignore[assignment]
-                    if vis.prompt == "[Decryption Failed]":
-                        vis.prompt = crypto.decrypt_text(str(vis.prompt), key=crypto.global_key)  # type: ignore[assignment]
+                    orig_prompt = str(vis.prompt)
+                    vis.prompt = crypto.decrypt_text(orig_prompt, key=umk)  # type: ignore[assignment]
+                    if vis.prompt == "[Decryption Failed]" and umk != crypto.global_key:
+                        vis.prompt = crypto.decrypt_text(orig_prompt, key=crypto.global_key)  # type: ignore[assignment]
             except Exception as e:
                 logger.error(f"Vision decryption failed: {e}")
                 vis.result = "[Fehler: Entschlüsselung fehlgeschlagen]"  # type: ignore[assignment]

db_ops.py

@@ -53,7 +53,7 @@ def load_api_key() -> str:
     if not api_key:
         print("❌ Error: GLADIA_API_KEY not found in environment or .env file")
         sys.exit(1)
-    debug(f"API key loaded: {api_key[:10]}...")
+    debug("API key loaded: present")
     return api_key
 
 

gladia.py

@@ -186,7 +186,7 @@ def main():
 
         print(f"[*] Downloading {len(to_download)} files...")
         for f in to_download:
-            path = os.path.join(target_dir, f["name"])
+            path = os.path.join(target_dir, os.path.basename(f["name"]))
             try:
                 downloader.download_stream(f, path)
             except Exception as e:

pcloud_dl.py

@@ -13,11 +13,11 @@ openai>=1.0.0
 # HTTP / networking
 requests>=2.32.0
 httpx>=0.27.0  # async HTTP for dictation_manager.process_crispasr_remote
-websockets
-certifi
-idna
-urllib3
-chardet
+websockets>=16.0
+certifi>=2026.5.20
+idna>=3.18
+urllib3>=2.7.0
+chardet>=7.4.3
 requests-toolbelt>=0.8.0
 
 # Database
@@ -43,10 +43,10 @@ python-pptx>=0.6.23
 lxml>=4.9.0
 
 # Misc
-python-dotenv
-beautifulsoup4
-pydub
-yt-dlp
+python-dotenv>=1.1.0
+beautifulsoup4>=4.14.3
+pydub>=0.25.1
+yt-dlp>=2026.3.17
 
 # ── ppt-llm presentation generator (pptx_tool.py / pptx_engine) ──
 pydantic>=2.0

requirements.txt

@@ -0,0 +1,53 @@
+"""Shared SSRF guard — validates URLs before server-side fetches.
+
+Rejects non-HTTP(S) schemes, resolves hostnames and blocks
+private / loopback / link-local / reserved / metadata addresses.
+"""
+
+import ipaddress
+import socket
+from urllib.parse import urlparse
+
+__all__ = ["is_safe_url"]
+
+
+def is_safe_url(url: str) -> tuple[bool, str]:
+    """Check whether *url* is safe to fetch server-side.
+
+    Returns ``(True, "")`` when safe, or ``(False, reason)`` when the
+    URL must be refused (private IP, bad scheme, etc.).
+    """
+    try:
+        p = urlparse(url)
+    except Exception as e:
+        return False, f"unparseable URL: {e}"
+
+    if p.scheme not in ("http", "https"):
+        return False, f"scheme not allowed: {p.scheme!r}"
+
+    host = p.hostname
+    if not host:
+        return False, "no hostname"
+
+    try:
+        infos = socket.getaddrinfo(host, None)
+    except Exception as e:
+        return False, f"DNS resolution failed: {e}"
+
+    for _fam, _t, _p, _c, sockaddr in infos:
+        ip_str = sockaddr[0]
+        try:
+            ip = ipaddress.ip_address(ip_str)
+        except ValueError:
+            return False, f"invalid resolved IP: {ip_str}"
+        if (
+            ip.is_private
+            or ip.is_loopback
+            or ip.is_link_local
+            or ip.is_reserved
+            or ip.is_multicast
+            or ip.is_unspecified
+        ):
+            return False, f"address blocked (private/loopback/metadata): {ip_str}"
+
+    return True, ""