Curly-Haired-Baboon

#2808of 53,632
89.8Total CVSS
Vulnerabilities · 10
High
3
Critical
7
PT-2026-39288
9.4
2026-05-08
Siyuan · Siyuan · CVE-2026-44670
**Name of the Vulnerable Software and Affected Versions** SiYuan versions prior to 3.7.0 **Description** The software fails to escape HTML when storing and rendering Attribute View (AV) names. The kernel stores these names without escaping and uses a raw string replacement to embed them in HTML before sending them to clients via WebSocket. This allows for HTML injection through three client paths: `render.ts` (via `outerHTML`), `Title.ts` (via `innerHTML`), and `transaction.ts` (via `innerHTML`). Because the main BrowserWindow is configured with `nodeIntegration: true`, `contextIsolation: false`, and `webSecurity: false`, this HTML injection can be escalated to Node.js code execution (RCE) on the victim's desktop. The payload is persistent, as it is stored on disk, replicates through sync transports (S3, WebDAV, cloud), and survives export-import processes. It can be triggered by any user role opening a document bound to the affected AV. Technical details include the use of the `/api/transactions` endpoint with the `setAttrViewName` action to plant the payload. The vulnerability is reachable via local-origin requests, including those from allowlisted `chrome-extension://` origins, allowing malicious browser extensions to execute the attack. **Recommendations** Update to version 3.7.0. As a temporary mitigation, restrict access to the `/api/transactions` endpoint or avoid importing `.sy.zip` files from untrusted sources.
PT-2026-39003
9.6
2026-05-08
Praisonai · Praisonai · CVE-2026-44336
**Name of the Vulnerable Software and Affected Versions** PraisonAI versions prior to 4.6.34 **Description** The Model Context Protocol (MCP) server in PraisonAI contains a path traversal flaw in its file-handling tools. The server registers four tools by default: 'praisonai.rules.create', 'praisonai.rules.show', 'praisonai.rules.delete', and 'praisonai.workflow.show'. These tools accept path or filename strings via the `tools/call` endpoint and join them to the `~/.praison/rules/` directory (or accept absolute paths in the case of 'praisonai.workflow.show') without performing containment checks. Additionally, the JSON-RPC dispatcher passes `params["arguments"]` to handlers via `**kwargs` without validating them against the input schema. An attacker can use the `rule name` variable with traversal sequences (e.g., `../../`) to read, write, or delete any file the running user has permissions for. This can be escalated to arbitrary code execution by writing a Python `.pth` file into the user site-packages directory, which is then executed by any subsequent Python process spawned by the user. This issue can be triggered via an MCP-connected LLM through indirect prompt injection (e.g., malicious web content), or via the `praisonai mcp serve --transport http-stream` endpoint, which defaults to no authentication on loopback. **Recommendations** Update to version 4.6.34. As a temporary workaround, restrict access to the loopback interface or ensure the `--api-key` flag is used when running the MCP server with HTTP-stream transport to prevent unauthenticated access.