Seankohjs

The `/credentials/signurl` endpoint incorrectly uses the `X-Rucio-Account` HTTP header to determine authorization instead of the account identity stored in the authenticated token. An authenticated user with low privileges can provide `root` as the value for the `X-Rucio-Account` header to bypass the `get signed url()` permission check, which is restricted to root users. This allows the attacker to obtain a presigned read, write, or delete URL for any arbitrary object using the stored S3, GCS, or Swift credentials of the Rucio server. Additionally, in multi-VO deployments, the `X-Rucio-VO` header can be used to access other tenants. This leads to unauthorized data disclosure, modification, deletion, and cross-tenant access on cloud-backed RSEs.

An algorithmic complexity issue exists in the inline link parser's `parse link text()` function. Unauthenticated attackers can cause CPU exhaustion by submitting Markdown containing numerous unbalanced '[' or '![' characters. This triggers repeated unbounded forward scans, which can tie up worker processes for several minutes.

An issue exists in the `trust` setting's prefix match within `copier/ settings.py` where the template URL is compared against a trusted prefix using `str.startswith` without path normalization. Because the URL is normalized during the actual fetch process (using `Path.resolve()` for local paths and libcurl dot-segment removal for HTTPS), a template reference containing `..` that textually starts with a trusted prefix can be granted trust. This allows an attacker-controlled template to execute `tasks`, `migrations`, or `jinja extensions` without the `--trust` prompt, leading to arbitrary command execution.

A cross-site scripting issue exists in the `to markdown()` function. The problem occurs because inline code spans do not recognize blank lines as block boundaries. This allows attackers to inject blank lines into code or pre element text to break the inline span, resulting in sanitized HTML being emitted unescaped and subsequently re-parsed as live Markdown by compliant renderers.

The Docker API server fails to apply Server-Side Request Forgery (SSRF) destination checks on the streaming path. A remote unauthenticated client can exploit this by calling the 'POST /crawl/stream' endpoint or the 'POST /crawl' endpoint with the `crawler config.stream` variable set to true. The `handle stream crawl request()` function passes seed URLs directly to the crawler without validation, allowing the server to fetch and stream response bodies from internal, private, or link-local addresses.