38,630 results · Feed
-
High 7.1
PraisonAI: Arbitrary file write via unsanitized `user_id` in `FileMemory.__init__()` — path traversal to any writable location
PraisonAI is a multi-agent teams system. Prior to praisonaiagents 1.6.58, the FileMemory constructor joins unsanitized user_id into self.user_path. A caller supplying ../ or path separators can escape the memory directory and write JSON data to arbitrary proc…
-
High 8.8
PraisonAI: `--api-key` flag on `praisonai serve` is not properly enforced
PraisonAI is a multi-agent teams system. Prior to praisonai 4.6.58, praisonai serve agents and praisonai serve unified parse --api-key but _create_agents_app() and _create_unified_app() do not install a credential check. Unauthenticated callers can reach POST…
-
Medium 6.8
PraisonAI: Server-Side Request Forgery via DNS rebinding bypass in webhook_url validation
PraisonAI is a multi-agent teams system. Prior to praisonai 4.6.58, the Jobs API validate_webhook_url() path fails open on socket.gaierror and does not bind the validated address to the later request. An attacker webhook_url can later resolve to 127.0.0.1, 16…
-
High 7.1
PraisonAI: Webhook SSRF via DNS fail-open in `JobSubmitRequest.validate_webhook_url()` — bypass of CVE-2026-40114
PraisonAI is a multi-agent teams system. Prior to praisonai 4.6.58, JobSubmitRequest.validate_webhook_url() accepts webhook_url when resolution raises socket.gaierror because the exception path uses except socket.gaierror: pass. JobExecutor._send_webhook() la…
-
High 7.3
PraisonAI: [Auth Bypass] `praisonai serve agents --api-key` is silently ignored — agent-invocation routes (`POST /agents`, `POST /agents/{agent_name}`) run unauthenticated
PraisonAI is a multi-agent teams system. Prior to praisonai 4.6.51, praisonai serve agents parses config["api_key"] but _create_agents_app() does not authenticate POST /agents or POST /agents/{agent_name}. Missing or incorrect bearer and X-API-Key values stil…
-
High 7.1
PraisonAI: [Path Traversal] agent tools escape the configured workspace via symlinks
PraisonAI is a multi-agent teams system. Prior to praisonai 4.6.51, is_path_within_directory() uses os.path.abspath() rather than os.path.realpath() for the workspace boundary. A symlink inside workspace can point outside and still pass the check, allowing re…
-
Medium 6.1
PraisonAI: ast_grep_rewrite rewrites arbitrary files without the @require_approval gate enforced on every sibling mutation tool
PraisonAI is a multi-agent teams system. Prior to praisonaiagents 1.6.58, ast_grep_rewrite lacks the @require_approval decorator used by sibling mutation tools. With dry_run=False, an agent-controlled call can pass --update-all and a broad path to rewrite mat…
-
High 7.5
PraisonAI: SSRF in web_crawl tool via redirect-following and DNS rebinding (validate-then-fetch gap)
PraisonAI is a multi-agent teams system. In versions prior to 1.6.58, the web_crawl tool performs its SSRF check only on the initially supplied URL, allowing the protection to be bypassed so the tool connects to attacker-chosen internal destinations. The chec…
-
High 8.6
PraisonAI serve agents --api-key is ignored, allowing unauthenticated remote agent execution
PraisonAI is a multi-agent teams system. From praisonai 4.6.34 until 4.6.58, praisonai serve agents accepts --api-key but _create_agents_app() does not authenticate POST /agents or POST /agents/{agent_name}. A network caller can invoke configured agents witho…
-
High 8.5
PraisonAI: SSRF protection bypass in `spider_tools._host_is_blocked()` via DNS-resolved hostnames (`127.0.0.1.nip.io`)
PraisonAI is a multi-agent teams system. Prior to praisonaiagents 1.6.58, spider_tools._host_is_blocked() does not resolve ordinary hostnames before scrape_page fetches them. A hostname such as 127.0.0.1.nip.io passes validation and resolves to loopback, perm…
-
Medium 6.5
PraisonAI: Unauthenticated unbounded session accumulation in the PraisonAI MCP HTTP server (memory exhaustion; session TTL never enforced)
PraisonAI is a multi-agent teams system. Prior to praisonai 4.6.58, the MCP HTTP Stream mcp_post handler creates a new _sessions entry for every initialize request but does not call _cleanup_sessions or enforce a maximum. An unauthenticated caller can exhaust…
-
High 8.2
praisonaiagents: AgentServer declares auth_token but never enforces it on any route (CWE-862)
PraisonAI is a multi-agent teams system. Prior to praisonaiagents 1.6.58, AgentServer exposes ServerConfig.auth_token but AgentServer._create_app does not check it on any route. A remote caller can subscribe, publish, and perform other actions without a valid…
-
Medium 6.9
PraisonAI: Origin validation bypass in MCP HTTP Stream transport allows browser-mediated unauthenticated tool execution on local MCP server
PraisonAI is a multi-agent teams system. Prior to praisonai 4.6.58, the MCP HTTP Stream _validate_origin method accepts request_origin.startswith(allowed), so the attacker-controlled localhost.evil.example HTTP origin matches the localhost allowlist. Without …
-
High 7.7
PraisonAI has a`web_crawl` SSRF protection bypass via unchecked redirect targets
PraisonAI is a multi-agent teams system. In versions 1.5.128 through 1.6.57, the praisonaiagents.tools.web_crawl_tools.web_crawl() function is vulnerable to server-side request forgery. While it validates the initially supplied URL and blocks direct loopback …
-
High 7.8
PraisonAI workflow include bypasses tools.py autoload opt-in and executes included recipe code
PraisonAI is a multi-agent teams system. In versions 3.9.26 through 4.6.57 of praiseonai and 0.12.12 through 1.6.57 of praiseonaiagents, the workflow "include" feature is vulnerable to code execution. Workflow._execute_include() implicitly imports and runs an…
-
High 7.5
PraisonAI: SSRF via redirect-following in praisonaiagents web_crawl
PraisonAI is a multi-agent teams system. Prior to praisonaiagents 1.6.58, the web_crawl function validates only the initial URL before _crawl_with_httpx uses httpx.Client(follow_redirects=True). Redirect targets are not revalidated, so an attacker who influen…
-
Medium 5.3
vLLM before 0.27.0 Denial of Service via DeepStream Backend
vLLM before 0.27.0 fails to properly classify DeepStream as a GPU backend and omits pixel-limit enforcement in its decode path. Unauthenticated attackers can activate DeepStream at request time to initialize the process-wide GPU decode pool and submit video t…
-
High 7.6
Purpose-limited JWT accepted as full bearer authentication in AshAuthentication
Improper Authentication vulnerability in team-alembic AshAuthentication allows purpose-limited JWTs to be replayed as full bearer API credentials when a resource uses stateless bearer-token verification. The bearer-token authentication helper AshAuthenticatio…
-
Low 2.1
Reflected XSS in AshAuthentication confirmation and magic link interaction forms
Improper Neutralization of Input During Web Page Generation (XSS) vulnerability in team-alembic AshAuthentication allows reflected cross-site scripting via the confirmation and magic link interaction forms. When a strategy is configured with require_interacti…
-
Medium —
vibeio-http has a DoS vulnerability in HTTP/1.x chunked encoding parser triggered by maliciously crafted chunk lengths
When using the affected versions of the `vibeio-http` crate, an attacker could craft a malicious HTTP/1.x request with a large chunk length (between `usize::MAX - 1` and `usize::MAX` inclusive) and send it, causing the server to crash (integer overflow panic …
-
Low 3.1
Cloudreve before 4.16.1 Path Traversal via Remote Download
Cloudreve before 4.16.1 fails to properly sanitize file paths returned by remote downloaders, allowing authenticated users to create files outside the selected destination directory. Attackers can exploit path traversal sequences in downloader metadata to wri…
-
Medium 5.3
Cloudreve before 4.16.1 Authentication Bypass via Cached Context Hint
Cloudreve before 4.16.1 fails to revalidate share access when restoring cached navigator state from a context_hint UUID. Attackers who previously had valid share access can replay the cached hint to generate signed file URLs for up to 300 seconds after the sh…
-
High 7.2
Authenticated Arbitrary File Write via Database Import and Xray Log Path Manipulation
3X-UI is a web control panel for managing Xray-core servers. Prior to 3.3.1, an authenticated administrator can abuse the database import functionality to achieve arbitrary file write on the host by modifying Xray configuration values stored in the database. …
-
Medium —
Gorilla WebSocket Uses Cryptographically Weak PRNG for WebSocket Mask Key
gorilla/websocket used `math/rand` (cryptographically weak pseudo-random number generator) to generate WebSocket frame mask keys prior to commit d67f4185. WebSocket masking keys MUST be unpredictable to prevent frame content injection attacks. math/rand produ…
-
Low —
netfoil vulnerable to improper handling of untrusted DoH response data
### Summary HTTPS RR data returned from the DoH server was not properly handled. Unverified APLN data (which could contain arbitrary byte sequences) was written directly to the log. It could also cause unnecessary memory usage. The single-byte response code w…
-
Medium 4.8
django CMS: Page cache ignores plugin-declared Vary headers (disclosure & poisoning)
django CMS is a content management system powered by Django. Prior to 5.0.8 and in 5.1.0a1, the django CMS page cache in cms/cache/page.py ignores request headers declared by plugins through get_vary_cache_on(). The _page_cache_key function includes the cache…
-
High 7.1
django CMS: Plugin move endpoint allows cyclic reparenting (DoS)
django CMS is an easy-to-use and developer-friendly enterprise content management system powered by Django. Prior to 5.0.8, the move_plugin endpoint in cms/admin/placeholderadmin.py accepts an attacker-controlled plugin_parent value without rejecting a plugin…
-
Medium —
tokio-postgres: Panic on a `DataRow` with fewer fields than columns allows denial of service
A malicious or compromised server can send a row containing fewer fields than its row description declares columns. Reading one of the missing columns then panics with an out-of-bounds index, aborting the calling task. This affects even the otherwise non-pani…
-
Medium —
postgres-protocol: Panic decoding a malformed `hstore` value allows denial of service
A malicious or compromised server can return a binary `hstore` value with an invalid internal length field, causing the client to panic while decoding it. Applications that connect only to a trusted database are not exposed; the risk applies to clients that m…
-
High —
postgres-protocol: Unbounded SCRAM iteration count allows a malicious server to cause CPU-exhaustion denial of service
A malicious, compromised, or man-in-the-middle server can supply an arbitrarily large SCRAM-SHA-256 PBKDF2 iteration count during authentication. The client runs it inline with no upper bound, pinning a `tokio` worker thread for minutes per connection, possib…
-
Medium 6.5
Sakai: IDOR in Profile Image Deletion Allows Any Authenticated User to Delete Any User's Profile Image
Sakai is a Collaboration and Learning Environment (CLE). From 23.0 until 23.5 and 25.3, the DELETE /api/users/{userId}/profile/image endpoint allows an authenticated user to delete another user's profile image because ProfileController.removeProfileImage() pa…
-
High 8.7
Sakai Conversations has a Stored XSS Issue
Sakai is a Collaboration and Learning Environment (CLE). From versions 23.0 to before 23.5, and versions 25.0 to before 25.3, the Sakai Conversations tool stores topic and post messages without HTML sanitization, and the frontend renders them using LitElement…
-
Critical 10.0
MCP Context Forge: RestrictedPython sandbox bypass via getattr builtin in python_sandbox_server
MCP Context Forge is an AI gateway, registry, and proxy for MCP, A2A, REST, and gRPC APIs. Prior to 1.0.2, the python_sandbox_server in mcp-servers/python/python_sandbox_server/src/python_sandbox_server/server_fastmcp.py exposes raw getattr through safe_built…
-
Critical 9.8CVE-2026-78329 Public exploit
Apache Camel: Camel-Undertow: the endpoint discarded the undertow-specific header filter strategy in favour of the base HTTP one, so the undertow filtering never ran on endpoint-configured routes
Improper input validation vulnerability in Apache Camel Undertow component. This issue affects Apache Camel: from 4.11.0 before 4.14.9, from 4.15.0 before 4.18.4, from 4.19.0 before 4.22.0. UndertowEndpoint defaulted its headerFilterStrategy field to the base…
-
Critical 9.1CVE-2026-66906 Public exploit
Apache Camel: Camel-Azure-Storage-Blob: the downloadBlobToFile operation built the local download target from the remote blob name without constraining it to the configured fileDir
Relative path traversal vulnerability in Apache Camel Azure Storage Blob component. This issue affects Apache Camel: from 4.0.0 before 4.14.9, from 4.15.0 before 4.18.4, from 4.19.0 before 4.22.0. The camel-azure-storage-blob component can download an Azure S…
-
High 7.5CVE-2026-66907 Public exploit
Apache Camel: Camel-Google-Storage: the consumer appended the remote object name to the configured downloadFileName directory without constraining the result
Relative path traversal vulnerability in Apache Camel Google Storage component. This issue affects Apache Camel: from 4.0.0 before 4.14.9, from 4.15.0 before 4.18.4, from 4.19.0 before 4.22.0. The camel-google-storage consumer downloads Google Cloud Storage o…
-
High 7.5CVE-2026-66908 Public exploit
Apache Camel: Camel-platform-http-main: when JWT authentication was configured with a keystore but no issuer or audience, the iss and aud claims were never validated, so any unexpired token signed by a trusted key was accepted
Improper Authentication vulnerability in Apache Camel Platform HTTP Main component. This issue affects Apache Camel: from 4.8.0 before 4.22.0. The camel-main embedded HTTP server can protect its endpoints with JWT authentication, configured through authentica…
-
Medium 6.5CVE-2026-59230 Public exploit
Apache Camel: Camel-Mail: the MimeMultipart data format copied MIME headers onto the Camel message without a header filter strategy when unmarshalling with headersInline enabled
Improper input validation vulnerability in Apache Camel. This issue affects Apache Camel: from 2.17.0 before 4.14.9, from 4.15.0 before 4.18.4, from 4.19.0 before 4.22.0. The camel-mail component ships a MimeMultipart data format that can unmarshal a MIME mul…
-
Medium 5.3CVE-2026-63621 Public exploit
Apache Camel: Camel-Knative: CloudEvent extension fields received in structured content mode were mapped onto message headers without applying any header filter strategy
Improper Input Validation, Improper Neutralization of Special Elements in Output Used by a Downstream Component ('Injection') vulnerability in Apache Camel Knative component The Knative consumer in camel-knative maps inbound CloudEvent attributes onto Camel m…
-
Critical 9.8CVE-2026-71300 Public exploit
Apache Camel: Camel-Atmosphere-Websocket: WebSocket dispatch header injection
Improper input validation vulnerability in Apache Camel Atmosphere Websocket component. This issue affects Apache Camel: from 4.0.0 before 4.14.9, from 4.15.0 before 4.18.4, from 4.19.0 before 4.22.0. The camel-atmosphere-websocket producer selects which conn…