Search: a-redirect.com
699 threats
Most recently published first
| Severity | CVSS | Signals | Threat | Source | Published | Score and star |
|---|---|---|---|---|---|---|
| Malicious | – | MaliciousGCVE Database The npm package 'backupsitetuff6' contains malicious code that abuses the npm registry by mass-publishing nearly 100 misleading package names. It includes a browser Service Worker that serves a heavily obfuscated web proxy frontend with a popunder redirect to an external site. The package does not execute code during installation in Node environments, so the primary harm is registry namespace abuse and browser-side malicious behavior. According to one source, any system with this package installed or running should be considered fully compromised, with immediate secret/key rotation recommended. | GCVE Database | |||
| Malicious | – | MaliciousGCVE Database The npm package 'ishowfeet1' contains malicious code that facilitates mass publication of nearly 100 similarly named packages by rewriting package names and republishing them. The package's main asset is a browser Service Worker that serves a heavily obfuscated web proxy frontend with a hidden popunder redirect to a malicious site. There is no automatic code execution during Node.js installation, so the threat is primarily registry abuse and malicious browser-side behavior when the assets are loaded. The package does not compromise the installer environment directly but poses a critical risk through its browser-based payload and namespace spamming. | GCVE Database | |||
| High | 8.4 | EPSS 0.3% | High8.4EPSS 0.3%CVE Database V5CVE-2026-61429 ## Summary The DNS-rebinding / redirect SSRF bypass in PRAI-05 is **not limited to the httpx/urllib backend**. When `crawl4ai` (headless Chromium via Playwright) is installed, `web_crawl` auto-selects `provider=crawl4ai`, and the headless browser re-resolves DNS and follows redirects on its own — with no per-connection SSRF guard. Runtime-confirmed read-back of the internal canary via both DNS rebinding and redirect (`provider: "crawl4ai"`). Status: runtime-confirmed addendum to PRAI-05. ## Details ### Affected component - Package `praisonaiagents` 4.6.63. Backend selected when `crawl4ai`+Playwright are installed. - Files: `src/praisonai-agents/praisonaiagents/tools/web_crawl_tools.py` (`_crawl_with_crawl4ai`) and `src/praisonai-agents/praisonaiagents/tools/crawl4ai_tools.py` (standalone crawl wrappers). ### Vulnerable code / root cause Path: `src/praisonai-agents/praisonaiagents/tools/web_crawl_tools.py` Function: `_crawl_with_crawl4ai` Snippet: ```python async with AsyncWebCrawler() as crawler: for url in urls: result = await crawler.arun(url=url) # headless browser: re-resolves DNS, follows redirects ``` Path: `src/praisonai-agents/praisonaiagents/tools/crawl4ai_tools.py` Function: `Crawl4AITools.crawl` / `crawl4ai` Snippet: ```python result = await crawler.arun(url=url, config=config) # no _is_safe_crawl_url / no per-connect validation ``` Issue: the only SSRF check is the single pre-fetch `_is_safe_crawl_url()` on the initial URL string in `web_crawl()` (see PRAI-05). The headless browser then resolves and connects independently and follows redirects in-browser — no resolved-IP pinning, no per-hop/per-connect validation. Input (`urls`) is attacker/agent-controlled; the sink is `crawler.arun(url=...)`; the guard is bypassed by DNS rebinding (TOCTOU) and by redirects (followed in-browser). ### Attack flow / Why bypassed / Security boundary Identical to PRAI-05: TOCTOU between the guard's resolution and the browser's connection; redirects followed in-browser; internal response returned as crawl `content`. See PRAI-05. ## Proof of Concept ### Environment `crawl4ai` + Playwright Chromium installed in a dedicated runtime container (`127.0.0.1:18081`), same controlled internal canary + rebinding DNS. Runnable assets: `PraisonAI-Runtime-Repro\runtime-files\` (`docker-compose.crawl.yml`). ### Steps to reproduce 1. `SSRF-04-01-Crawl4AI-DNS-Rebind-Trigger` → `127.0.0.1:18081`: ```http POST /tool/web_crawl HTTP/1.1 Host: 127.0.0.1:18081 Content-Type: application/json {"url":"http://rebind.lab:8081/secret"} ``` 2. Redirect variant `SSRF-04-03-Crawl4AI-Redirect-Readback`: `{"url":"http://redirector:8082/redirect-to-internal"}`. ### Expected result The crawl backend refuses internal destinations regardless of DNS timing/redirects. ### Actual result HTTP 200, `"provider":"crawl4ai"`, response `content` contains `PRAISONAI_INTERNAL_SECRET_CANARY_7f3a91` + `FAKE_INTERNAL_TOKEN_DO_NOT_USE_7f3a91` for both the DNS-rebinding and the redirect payload. ### Screenshots **Crawl4AI DNS rebinding read-back** The attacker-controlled `rebind.lab` URL is accepted by the Crawl4AI-backed `web_crawl` endpoint. PraisonAI returns the internal canary response body containing `PRAISONAI_INTERNAL_SECRET_CANARY_7f3a91`. <img width="1542" height="763" alt="01-Crawl4AI-Burp-Readback" src="https://github.com/user-attachments/assets/a03b3a1c-e382-4f67-8ea6-b766dceb4a69" /> **Crawl4AI DNS rebinding runtime evidence** The runtime log shows the Crawl4AI/Chromium backend resolving `rebind.lab` to an internal Docker IP during the fetch phase. The internal canary receives `GET /secret` from a headless Chrome user agent. <img width="1659" height="946" alt="02-Crawl4AI-DNS-Log-And-Internal-Hit" src="https://github.com/user-attachments/assets/d93db090-3e39-43d0-af3b-5d1d8f614db8" /> **Crawl4AI redirect read-back** The attacker-controlled redirector URL is accepted by the Crawl4AI-backed endpoint. PraisonAI follows the redirect and returns the internal canary response body. <img width="1544" height="771" alt="03-Crawl4AI-Redirect-Readback" src="https://github.com/user-attachments/assets/7d05c8aa-5bda-45bc-b94c-bef0bdcf055c" /> **Crawl4AI redirect runtime evidence** The controlled redirector returns `302 -> http://internal-canary:8081/secret`, and the internal canary receives `GET /secret` from the Crawl4AI/Chromium backend. <img width="1590" height="920" alt="04-Crawl4AI-Redirect-Internal-Hit-Log" src="https://github.com/user-attachments/assets/dd9c0fec-3583-43ab-b83c-4470fa6aa409" /> ## Impact Same class as PRAI-05: read-back SSRF to internal/metadata services, now on the crawl4ai backend. Broadens the affected surface (the flaw is in the shared validate-without-pinning design, not one backend). ## Suggested remediation Resolve-once + IP-pin + per-connect validation must also cover the crawl4ai backend (constrain the headless browser to the validated IP, or allowlist crawl destinations). | CVE Database V5 | ||
| Medium | 5.9 | EPSS 0.2% | Medium5.9EPSS 0.2%GCVE Database GNU Wget has a vulnerability in its FTP passive mode where it does not validate the IP address returned by the FTP PASV command. This flaw allows a malicious FTP server or an HTTP server redirecting to an FTP URL to redirect Wget's data connection to an arbitrary IP address and port. Exploiting this can enable server-side request forgery (SSRF), potentially allowing access to localhost or internal network services from the machine running Wget. | GCVE Database | ||
| Critical | 9.3 | EPSS 0.5% | CVE-2026-54072Authorizer · Unvalidated redirect_uri in /authorize leaks OAuth2 tokens to attacker-controlled URL Critical9.3EPSS 0.5%GCVE DatabaseCVE-2026-54072 The /authorize endpoint in github.com/authorizerdev/authorizer does not validate the redirect_uri parameter against allowed origins, allowing an attacker to supply a malicious redirect URI. When response_type=token or id_token, the server appends OAuth2 tokens as query parameters and redirects the user to the attacker-controlled URL. An unauthenticated attacker can obtain the client_id from a public GraphQL endpoint. This enables token theft if a logged-in user clicks a crafted link, allowing the attacker to impersonate the user for the token lifetime. | GCVE Database | ||
| Critical | 9.3 | EPSS 2.4% | CVE-2026-61459Flux159 mcp-server-kubernetes · Argument Injection Critical9.3EPSS 2.4%CVE Database V5CVE-2026-61459 MCP Server Kubernetes before 3.9.0 contains an argument injection vulnerability in structured tools (kubectl_get, kubectl_describe, kubectl_delete) that allows attackers to bypass the assertNoDangerousFlags security check by supplying resourceType and name parameters with leading dashes. Attackers can inject the --server flag to redirect kubectl commands to an attacker-controlled API server, causing the operator's bearer token to be transmitted externally and enabling full cluster compromise. EPSS 2.4%Vulnerability | CVE Database V5 | ||
| High | – | EPSS 0.2% | HighEPSS 0.2%GCVE DatabaseCVE-2026-6733 A security advisory from Red Hat addresses vulnerabilities in Red Hat Hardened Images RPMs, specifically involving the nodejs26 package and related components. One notable flaw is in the undici library's cookie parser, which incorrectly decodes cookie values, enabling HTTP response header injection. This can lead to session fixation, open redirect, or cache poisoning attacks. The advisory lists updated RPM packages to address these issues but does not specify fixed versions or detailed patch availability. No known exploits in the wild have been reported. | GCVE Database | ||
| Malicious | – | MaliciousGCVE Database --- _-= Per source details. Do not edit below this line.=-_ ## Source: amazon-inspector (c972c885927d861e92a964b7ab752e5e91ae2731092a51b08a64fbf97f015cdf) The package's main entry (px.js) is a browser-only script that creates a full-viewport iframe pointing at the hardcoded external URL https://mfmz.ywdyxn.cn/H.html?c=0gjr and appends it to document.body, hijacking any web page that loads this script via a CDN such as jsDelivr. The package ships no legitimate library API and would throw if require()'d in Node. The tarball also includes update_px8my.sh, an operator/rotation script that regex-replaces the redirect domain inside px.js, auto-bumps the patch version, runs npm publish, and then hits https://purge.jsdelivr.net/npm/px8my/px.js to force-refresh the CDN cache; a comment in that script notes that curl is blocked by the npm anti-abuse tool tirith. The combination — a hardcoded redirect payload plus a shipped domain-rotation/CDN-purge automation with an explicit anti-abuse-evasion note — establishes the package as a purpose-built malicious distribution vehicle abusing npm and jsDelivr to deliver browser redirects to end users of any site including this script. ## Source: ghsa-malware (78ae5dc7b5995ac8381a5716c7733e93d7de2c6500063fba58c05085cd03999d) Any computer that has this package installed or running should be considered fully compromised. All secrets and keys stored on that computer should be rotated immediately from a different computer. The package should be removed, but as full control of the computer may have been given to an outside entity, there is no guarantee that removing the package will remove all malicious software resulting from installing it. | GCVE Database | |||
| Medium | 6.8 | Medium6.8GCVE Database ### Impact > DefaultHttpClient follows redirects and forwards Authorization, Cookie, and Proxy-Authorization headers to redirect targets across domain boundaries. The blocklist only filters Host/Connection/TE/CT/CL. > Additionally, no maximum redirect count exists, enabling infinite loop DoS. > Affected: DefaultHttpClient.java lines 231-245, 1591, 2071 > Suggested fix: Strip sensitive headers on cross-domain redirects ### Patches It has been patched for versions: For Micronaut 5, versions equal or greater than 5.0.1 >= For Micronaut 4, versions equal or greater than 4.10.24 >= For Micronaut 3, versions equal or greater than 3.10.6 >= ### Workarounds No ### References Micronaut 5 Patch: https://github.com/micronaut-projects/micronaut-core/commit/9770328999f490bdfbb9e25addd45bf73d4a173a Micronaut 4 Patch: https://github.com/micronaut-projects/micronaut-core/commit/70cab4b44fbf985faba2846091f2356b5bd70719 Micronaut 3 Patch: https://github.com/micronaut-projects/micronaut-core/commit/64e539736b8168f201d868b02ace50fe14f57418 | GCVE Database | |||
| High | 7.5 | High7.5GCVE Database The Netty-based Micronaut HTTP Client does not impose a limit on HTTP redirections, potentially allowing an infinite redirect loop that could lead to a denial-of-service attack. ### Patches The following versions are patched: - For Micronaut 5, versions equal or greater than [5.0.1](https://github.com/micronaut-projects/micronaut-core/releases/v5.0.1) >= - For Micronaut 4, versions equal or greater than [4.10.24](https://github.com/micronaut-projects/micronaut-core/releases/v4.10.24) >= - For Micronaut 3, versions equal or greater than [3.10.7](https://github.com/micronaut-projects/micronaut-core/releases/v3.10.7) >= ### Workarounds No ### Resources Micronaut 5 Patch: https://github.com/micronaut-projects/micronaut-core/commit/6e88a972718d6e1521c5b3bb7766451798dba4e3 Micronaut 4 Patch: https://github.com/micronaut-projects/micronaut-core/commit/f1dffffec8fb5e3b7e94ae907ce0be3831e499d4 Micronaut 3 Patch: https://github.com/micronaut-projects/micronaut-core/commit/c06a2715ca7f78321bc3ca05f41cca78cd351320 | GCVE Database | |||
| Medium | 4.9 | EPSS 0.3% | Medium4.9EPSS 0.3%GCVE DatabaseCVE-2026-48737 ## Summary `is_global_address` in [`src/pyload/core/utils/web/check.py`](https://github.com/pyload/pyload/blob/1b12dc7f348db8c144e0f39215680415e90ca4d2/src/pyload/core/utils/web/check.py) is the central guard against SSRF-style outbound connections in pyload-ng. It tests whether a given IP is "globally routable" via Python's `ipaddress.ip_address(value).is_global`, and callers treat `not is_global` as "deny": ```python def is_global_address(value): try: return ipaddress.ip_address(value).is_global except ValueError: return False def is_global_host(value): ips = host_to_ip(value) return ips and all((is_global_address(ip) for ip in ips)) ``` Python's `ipaddress.IPv6Address.is_global` classifies the NAT64 well-known prefix as **globally routable** on every supported Python version (3.9 through 3.14 confirmed), and on older Pythons (3.9-3.11) the 6to4 prefix as well: | address | `is_global` on Py 3.9-3.11 | `is_global` on Py 3.12+ | wrapped IPv4 | |----------------------------------|---------------------------|--------------------------|----------------------| | `2002:7f00:0001::` (6to4) | True | False | 127.0.0.1 | | `2002:0a00:0001::` (6to4) | True | False | 10.0.0.1 | | `2002:a9fe:a9fe::` (6to4) | True | False | 169.254.169.254 (IMDS)| | `64:ff9b::a9fe:a9fe` (NAT64) | True | True | 169.254.169.254 | | `64:ff9b::7f00:1` (NAT64) | True | True | 127.0.0.1 | pyload-ng declares `python_requires = >=3.9` (`setup.cfg`), so deployments on Python 3.9-3.11 see the 6to4 path too. The NAT64 path is universal. `is_global` returns True for these wrappers, so `is_global_address` returns True and the deny check passes. The pycurl `PREREQFUNC` at [`src/pyload/core/network/http/http_request.py:680`](https://github.com/pyload/pyload/blob/1b12dc7f348db8c144e0f39215680415e90ca4d2/src/pyload/core/network/http/http_request.py#L680) consults the same helper just before TCP-connect: ```python if not self.allow_private_ip: is_proxy_ip = self.http_proxy_host and self.http_proxy_host == (conn_primary_ip, conn_primary_port) if not is_global_address(conn_primary_ip) and not is_proxy_ip: return pycurl.PREREQFUNC_ABORT return pycurl.PREREQFUNC_OK ``` On a host with 6to4 routing (legacy operator tunnels; `2002::/16` still configurable) or NAT64 (cloud IPv6-only subnets with NAT64 gateway), the encoded form routes to the embedded IPv4 and the curl connection terminates at the internal endpoint, defeating the deny. `is_global_host` (the helper that callers like `parse_urls` use against a URL hostname) feeds through `host_to_ip` which pins `family=AF_INET`, so hostname-based reach to these forms relies on the attacker supplying an IPv6 literal in the URL — but the curl PREREQFUNC sees the actual resolved IP (the AAAA returned for the hostname), so a hostname with an AAAA record set to one of the bypass forms reaches the same gap. Cross-reference: this is the same incomplete-coverage class as pydantic-ai's [GHSA-cqp8-fcvh-x7r3](https://github.com/pydantic/pydantic-ai/security/advisories/GHSA-cqp8-fcvh-x7r3) / CVE-2026-46678. pyload-ng's prior SSRF advisories [GHSA-7gvf-3w72-p2pg](https://github.com/pyload/pyload/security/advisories/GHSA-7gvf-3w72-p2pg) and [GHSA-8rp3-xc6w-5qp5](https://github.com/pyload/pyload/security/advisories/GHSA-8rp3-xc6w-5qp5) both went through `is_global_host` / `is_global_address`; the IPv6 transition gap is orthogonal to those redirect-bypass classes. ## Severity **MEDIUM** — `CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:C/C:L/I:N/A:L` = **4.7** - `AC:H` — exploitation requires the host network to route 6to4 (`2002::/16` traffic), have a NAT64 gateway, or otherwise resolve the IPv6 transition form to an internal IPv4 endpoint at the TCP layer. - `PR:L` — `parse_urls` ([`src/pyload/core/api/__init__.py:582`](https://github.com/pyload/pyload/blob/1b12dc7f348db8c144e0f39215680415e90ca4d2/src/pyload/core/api/__init__.py#L582)) requires `Perms.ADD`, which any account capable of adding links holds. The curl PREREQFUNC at [`http_request.py:680`](https://github.com/pyload/pyload/blob/1b12dc7f348db8c144e0f39215680415e90ca4d2/http_request.py#L680) is reached by every downloader plugin that runs after `is_global_host` passed. - `C:L/A:L` — internal-network recon and timing-based confirmation; cloud-metadata exfiltration on networks where the transition form actually routes. **CWE-918**: Server-Side Request Forgery (SSRF). ## Affected versions `pyload-ng` from the introduction of `is_global_address` / `is_global_host` in [`src/pyload/core/utils/web/check.py`](https://github.com/pyload/pyload/blob/1b12dc7f348db8c144e0f39215680415e90ca4d2/src/pyload/core/utils/web/check.py) up to an | GCVE Database | ||
| Critical | 9.6 | EPSS 0.4% | CVE-2026-15062Snowflake Snowpark Python SDK · SQL Injection Critical9.6EPSS 0.4%CVE Database V5CVE-2026-15062 SQL injection vulnerabilities in the Snowflake Snowpark Python SDK (snowpark-python) versions prior to 1.53.0 could allow authenticated low-privilege users to execute SQL beyond their authorization scope. An attacker could exploit these vulnerabilities by embedding SQL payloads in source database column names to escalate privileges via the DataFrameReader.dbapi() API by supplying a specially crafted location parameter to DataFrameWriter write methods to redirect a COPY INTO to an arbitrary source query, or by including a backslash-single-quote sequence in an export path to defeat the normalize_path() sanitizer and inject SQL via DataFrame.to_csv(). Successful exploitation may result in source database compromise, unauthorized cross-tenant data exfiltration, or unauthorized read of Snowflake account data. | CVE Database V5 | ||
| Malicious | – | MaliciousGCVE Database --- _-= Per source details. Do not edit below this line.=-_ ## Source: amazon-inspector (36241e999d92bad738082bf420fc144391735f9f3707bc4baa0153b4cb0eec8a) gas-log impersonates the legitimate eth-gas-reporter package: it reuses that project's README, keywords, badges, and CHANGELOG, and the README even instructs users to install the target name. The exported Mocha reporter in index.js contains two near-identical implementations. The real reporter (Gas) is defined but not exported; the exported function (log) sets `var opt = 1` and unconditionally enters an else branch that calls `utils.connectNet(...)`, making the always-false `if (!opt)` guard a dead-code gate that hides the payload as an unreachable variant of the legitimate code. utils.connectNet spawns a detached, unref'd `node lib/syncResolve.js` child so the loader outlives the parent Mocha process. lib/syncResolve.js issues `axios.get(DEV_URI, { headers: { 'x-secret-key': '_' } })`, extracts `.data.Cookie` from the response, and passes it to `new Function.constructor('require', result)` with `require` bound in, then invokes it — an eval of attacker-controlled response bytes with full Node module access, retried 5 times. DEV_URI is empty in this shipped version so the fetch will fail today, but the mechanism is complete and armed: any future publish, DNS resolution, or same-origin redirect activates full RCE on any developer who runs mocha with this reporter. The typosquat framing, the dead-code gate, and the fetch → eval → detached-child chain together are unambiguous supply-chain attack intent. ## Source: ghsa-malware (30475d80d69fa45b40a86bc3d91a9ed3c29f15d9c6d500760fc91e1ea303e8d5) Any computer that has this package installed or running should be considered fully compromised. All secrets and keys stored on that computer should be rotated immediately from a different computer. The package should be removed, but as full control of the computer may have been given to an outside entity, there is no guarantee that removing the package will remove all malicious software resulting from installing it. | GCVE Database | |||
| Medium | 5.9 | EPSS 0.4% | CVE-2026-53572V2 · KEDA has PostgreSQL connection string parameter injection via incomplete whitespace escaping Medium5.9EPSS 0.4%GCVE DatabaseCVE-2026-53572 KEDA versions prior to 2.20.0 have a vulnerability in the PostgreSQL scaler connection string construction where tenant-controlled parameters are insufficiently escaped. This allows injection of additional connection parameters by including unescaped ASCII whitespace characters other than space, such as tabs, which can alter connection behavior. An attacker with the ability to create TriggerAuthentication or ScaledObject resources can downgrade TLS connections, redirect connections to attacker-controlled hosts, and inject arbitrary libpq parameters. | GCVE Database | ||
| High | 8.3 | High8.3GCVE Database ### Summary `agentConn.apiClient()` used the default redirect behavior of `http.Client` while its custom transport dialed the host from the request URL as long as the port was the workspace agent HTTP API port (`4`). Agent tailnet IPs are deterministic from agent UUIDs, so a malicious workspace agent could return a redirect to another agent's tailnet IP and cause the control-plane client to send the follow-up workspace agent API request to the victim agent instead of the intended agent. Redirects that preserve the method and body, such as HTTP 307 and 308, allow replay of write and process-start requests. HTTP 301, 302 and 303 redirects only convert POST requests to GET requests, so they are sufficient to demonstrate redirected reads but not write or command-execution primitives. ### Impact An authenticated user with a running workspace and control of a modified workspace agent can redirect workspace agent API requests made to their agent toward another online workspace agent reachable on the tailnet. If the attacker knows the victim agent UUID, they can derive the victim's tailnet IP and target the victim's file APIs to read or write files as the victim workspace user. In affected versions that expose the workspace agent process API, the same primitive can be chained to command execution as the victim workspace user by writing a payload through the redirected file API and then redirecting a process-start request to execute it. This crosses workspace and tenant boundaries. ### Patches The fix disables automatic redirect following for workspace agent API clients and pins all workspace agent API dials to the intended agent's deterministic tailnet address. Requests whose URL host does not match the intended agent are rejected instead of being dialed. The fix was backported to all supported release lines: | Release line | Patched version | |---|---| | 2.34 | [v2.34.4](https://github.com/coder/coder/releases/tag/v2.34.4) | | 2.33 | [v2.33.10](https://github.com/coder/coder/releases/tag/v2.33.10) | | 2.32 | [v2.32.9](https://github.com/coder/coder/releases/tag/v2.32.9) | | 2.29 (ESR) | [v2.29.19](https://github.com/coder/coder/releases/tag/v2.29.19) | ### Workarounds None. Upgrading is required. ### Resources - Fix: #26600 | GCVE Database | |||
| Critical | 10.0 | Critical10.0GCVE Database --- title: Unauthenticated CRUD on /api/providers and Full API Key Leak via /api/usage/stats product: 9Router version: <= 0.4.41 severity: critical cve_request: true --- ## Summary Multiple critical API security vulnerabilities were discovered in 9Router's Next.js dashboard. The `/api/providers` endpoints lack authentication entirely, allowing anyone to create, read, update, and delete provider connections. Additionally, `/api/usage/stats` exposes full plaintext API keys, and `/api/usage/request-logs` + `/api/usage/request-details` expose all users' request history and full conversation contents (including system prompts, user messages, assistant responses) without authentication. ## Affected Endpoints | Endpoint | Method | Issue | |---|---|---| | `/api/providers` | GET | Lists all provider connections with partial credentials, OAuth tokens, account IDs | | `/api/providers/:id` | GET | Read any single provider detail (IDOR) | | `/api/providers` | POST | Create arbitrary provider connections with attacker-controlled API keys | | `/api/providers/:id` | PUT | Modify any existing provider connection | | `/api/providers/:id` | DELETE | Delete any provider connection | | `/api/usage/stats` | GET | Exposes full plaintext API keys, per-account usage breakdown, cost data | | `/api/usage/request-logs` | GET | Exposes all users' request logs (model, tokens, cost, timestamp, provider) | | `/api/usage/request-details/:id` | GET | Exposes full conversation turns including system prompts, user messages, assistant responses | | `/api/version` | GET | Exposes current version info | | `/api/models` | GET | Exposes full model routing catalog | | `/api/v1/models` | GET | Exposes model list | ## Impact ### Critical: Provider CRUD without authentication An attacker can: 1. **Add a malicious provider** — inject a provider that proxies through their server, capturing all prompts, responses, and API keys routed through 9Router 2. **Modify existing providers** — replace API keys with attacker-controlled ones, redirect traffic 3. **Delete all providers** — cause complete denial of service 4. **Read all provider configurations** — harvest partial credentials, GitHub Copilot OAuth tokens, Cloudflare account IDs, email addresses ### Critical: Full API key leak via /api/usage/stats The endpoint returns complete API key strings (e.g., `sk-...`) in plaintext alongside usage data per key, enabling unauthorized use of connected AI provider accounts. ### Critical: Conversation history leak `/api/usage/request-details` returns the full conversation history of other users' AI sessions, including system prompts, user messages, assistant responses, tool calls, and reasoning traces. ## Steps to Reproduce ### 1. Unauthenticated read of all providers ```bash curl -s https://<host>/api/providers ``` Returns all provider connections with email addresses, auth type, account IDs, and partial API key prefixes. ### 2. Create a provider without authentication ```bash curl -X POST https://<host>/api/providers \ -H "Content-Type: application/json" \ -d '{"provider":"openai","authType":"apikey","name":"rogue","apiKey":"sk-attacker-controlled"}' ``` Returns the created connection object with a new UUID and `isActive: true`. ### 3. Modify an existing provider without authentication ```bash curl -X PUT https://<host>/api/providers/<existing-uuid> \ -H "Content-Type: application/json" \ -d '{"name":"modified","apiKey":"sk-attacker-key"}' ``` Returns the updated connection object. ### 4. Delete a provider without authentication ```bash curl -X DELETE https://<host>/api/providers/<existing-uuid> ``` Returns `{"message":"Connection deleted successfully"}`. ### 5. Read full usage stats with API keys ```bash curl -s https://<host>/api/usage/stats ``` Returns full API key strings, per-account token/cost breakdown, recent requests. ### 6. Read request logs ```bash curl -s "https://<host>/api/usage/request-logs?page=1&pageSize=50" ``` Returns paginated request logs with timestamps, models, providers, user emails, token counts. ### 7. Read full conversation ```bash curl -s https://<host>/api/usage/request-details/<request-uuid> ``` Returns complete conversation turns for that request. ### 8. Read version info ```bash curl -s https://<host>/api/version ``` Returns `{"currentVersion":"0.4.19","latestVersion":"0.4.45","hasUpdate":true}`. ## Root Cause The Next.js API routes under `src/app/api/*` lack authentication middleware on several endpoints. Specifically: - `/api/providers/*` — No auth check before CRUD operations on provider connections stored in the database - `/api/usage/stats` — No auth check before returning aggregated usage data including full API keys - `/api/usage/request-logs` — No auth check before returning request history - `/api/usage/request-details/:id` — No auth check before returning full conversation contents ## Suggested Fix 1. Add authentication middleware to all `/api/providers/*` routes (GET, POST, | GCVE Database | |||
| Medium | 5.4 | EPSS 0.3% | CVE-2026-55437V2 · Coder vulnerable to stored HTML injection via workspace agent logs in AgentLogLine component Medium5.4EPSS 0.3%GCVE DatabaseCVE-2026-55437 ### Summary The `AgentLogLine` dashboard component instantiated `ansi-to-html` without `escapeXML: true` and inserted the result via `dangerouslySetInnerHTML` so HTML embedded in workspace agent log lines was rendered as live markup. Server-side sanitization did not neutralize HTML metacharacters. > **Note:** Exploitation requires a victim to view attacker-controlled agent logs in the dashboard. ### Impact A user who could run a workspace could emit arbitrary HTML into agent logs; when another user, including an administrator, viewed the workspace page, it rendered in their session. Content Security Policy blocked inline scripts but an attacker could still inject a `meta refresh` redirect, `style` rules for UI redressing or CSS-based exfiltration or external `img` beacons. This required workspace-owner access and a victim viewing the page. ### Patches The fix enables `escapeXML: true` so HTML metacharacters are escaped before DOM insertion. The fix was backported to all supported release lines: | Release line | Patched version | |---|---| | 2.34 | [v2.34.2](https://github.com/coder/coder/releases/tag/v2.34.2) | | 2.33 | [v2.33.8](https://github.com/coder/coder/releases/tag/v2.33.8) | | 2.32 | [v2.32.7](https://github.com/coder/coder/releases/tag/v2.32.7) | | 2.29 (ESR) | [v2.29.17](https://github.com/coder/coder/releases/tag/v2.29.17) | ### Workarounds None. ### Resources - Fix: #25808 ### Credits Coder would like to thank Anthropic's Security Team (ANT-2026-22449) for independently disclosing this issue! | GCVE Database | ||
| Medium | 6.5 | EPSS 0.7% | CVE-2026-49097Apache Camel · Improper Input Validation Medium6.5EPSS 0.7%CVE Database V5CVE-2026-49097 Improper Input Validation, Improper Neutralization of Special Elements in Output Used by a Downstream Component ('Injection') vulnerability in Apache Camel IRC component. The camel-irc producer chooses the destination of an outgoing IRC message from the irc.sendTo Exchange header (the constant IrcConstants.IRC_SEND_TO, value irc.sendTo); when that header is present it overrides the channel list configured on the endpoint, and the message is sent only to the specified destination. This and the component's other control headers (irc.target, irc.messageType, irc.user.*, irc.num, irc.value) used plain, non-Camel-prefixed values. Because these names do not start with the Camel / camel prefix, HttpHeaderFilterStrategy - which blocks only the Camel header namespace on the HTTP boundary - let them pass from an inbound HTTP request straight into the Exchange. In a route that bridges an HTTP consumer (for example platform-http) into an irc: producer, any HTTP client could therefore set the irc.sendTo header and redirect a message that the route intended for a configured channel to an arbitrary IRC channel or user - exfiltrating the message content to an attacker-chosen nickname, leaking it into a public channel, or delivering messages that appear to come from the bot. No credentials are required when the bridging consumer is unauthenticated. This issue affects Apache Camel: from 4.0.0 before 4.14.8, from 4.15.0 before 4.18.3, from 4.19.0 before 4.21.0. Users are recommended to upgrade to version 4.21.0, which fixes the issue. If users are on the 4.14.x LTS releases stream, then they are suggested to upgrade to 4.14.8. If users are on the 4.18.x releases stream, then they are suggested to upgrade to 4.18.3. After upgrading, routes that set IRC headers via the raw header names must use the CamelIrc* names (for example CamelIrcSendTo) instead of the old irc.* values. For deployments that cannot upgrade immediately, strip the irc.* headers from any untrusted ingress before the irc: producer (for example removeHeaders('irc.*') at the start of the route), and set the IRC destination from a trusted source. | CVE Database V5 | ||
| Low | 3.7 | EPSS 0.6% | CVE-2026-46584Apache Camel Mail · Improper Input Validation Low3.7EPSS 0.6%CVE Database V5CVE-2026-46584 Improper Input Validation, Exposure of Sensitive Information to an Unauthorized Actor vulnerability in Apache Camel Mail Component. The camel-mail producer (MailProducer.getSender) scanned the outgoing Exchange for message headers in the mail.smtp. / mail.smtps. namespace and, when any were present, built a per-message JavaMail sender with those values applied as JavaMail session properties, overriding the endpoint configuration. This namespace is Camel-internal - only MailProducer interprets it - and was not blocked by any HeaderFilterStrategy, so the values could originate from any inbound protocol (for example platform-http query parameters or request headers, or JMS / Kafka messages from untrusted producers) that feeds a route ending in an smtp / smtps producer without an intervening removeHeaders. The maximal impact is version-dependent: on releases before 4.19.0, setting mail.smtp.host redirects the SMTP connection to a server under the attacker's control, and because the producer then authenticates with the endpoint's configured username and password those credentials are transmitted to the attacker; on 4.19.0 and later the producer connects to the endpoint's configured host explicitly, so the reachable impact is limited to weakening transport security (for example mail.smtp.ssl.trust, mail.smtp.starttls.enable or mail.smtp.socks.host) and interception of the outgoing message rather than host redirect. Exploitation requires a route that channels untrusted input into the mail producer without stripping the namespace. This issue affects Apache Camel: from 4.0.0 before 4.14.8, from 4.15.0 before 4.18.3, from 4.19.0 before 4.21.0. Users are recommended to upgrade to version 4.21.0, which fixes the issue. If users are on the 4.14.x LTS releases stream, then they are suggested to upgrade to 4.14.8. If users are on the 4.18.x releases stream, then they are suggested to upgrade to 4.18.3. After upgrading, the per-message override is disabled by default; enable it only on trusted endpoints with useJavaMailSessionPropertiesFromHeaders=true. For deployments that cannot upgrade immediately, strip the namespace before the mail producer with removeHeaders('mail.smtp.*') and removeHeaders('mail.smtps.*') between any untrusted ingress and the smtp / smtps producer. Even with the opt-in enabled, route authors should still strip the namespace on any path that carries untrusted input. | CVE Database V5 | ||
| High | 8.1 | EPSS 0.2% | High8.1EPSS 0.2%CVE Database V5CVE-2026-12740 Plack::Middleware::OAuth versions through 0.10 for Perl do not support the OAuth 2.0 state parameter. RequestTokenV2 builds the provider authorization redirect without issuing a state value, and AccessTokenV2 exchanges the callback code and registers the resulting token into the session (register_session) without verifying that the callback corresponds to an authorization request this session initiated. Any application that uses this middleware for OAuth 2.0 login is exposed to login cross-site request forgery: because the callback is not bound to the session that began the flow, an attacker who starts an authorization with their own provider account can deliver the resulting callback to a victim, causing the victim's session to complete the attacker's authorization and associating the attacker's provider identity and access token with that session. Where the application persists this as an account link, the attacker may retain access to the victim's account through their own provider credentials. | CVE Database V5 |
Page 24 of 43