FastFeedParser has an infinite redirect loop DoS via meta-refresh chain

Description

Summary

When parse() fetches a URL that returns an HTML page containing a <meta http-equiv="refresh"> tag, it recursively calls itself with the redirect URL — with no depth limit, no visited-URL deduplication, and no redirect count cap. An attacker-controlled server that returns an infinite chain of HTML meta-refresh responses causes unbounded recursion, exhausting the Python call stack and crashing the process. This vulnerability can also be chained with the companion SSRF issue to reach internal network targets after bypassing the initial URL check.

Details

parse() catches ValueError on XML parse failure, extracts a meta-refresh URL from the HTML response via _extract_meta_refresh_url(), and tail-calls itself with that URL. The recursive call is unconditional — there is no maximum redirect depth, no set of already-visited URLs, and no guard against self-referential or looping redirects.

fastfeedparser/main.pyparse() (recursive sink):

def parse(source: str | bytes, ...) -> FastFeedParserDict:
    is_url = isinstance(source, str) and source.startswith(("http://", "https://"))
    if is_url:
        content = _fetch_url_content(source)
    try:
        return _parse_content(content, ...)
    except ValueError as e:
        ...
        redirect_url = _extract_meta_refresh_url(content, source)
        if redirect_url is None:
            raise
        return parse(redirect_url, ...)   # ← unconditional recursion, no depth limit

_extract_meta_refresh_url() uses urljoin(base_url, match.group(1)) so relative, protocol-relative (//host/path), and absolute URLs in the content= attribute are all followed.

PoC

No live server required. The following monkeypatches _fetch_url_content to return an infinite HTML meta-refresh chain and confirms unbounded recursion:

import fastfeedparser.main as m

call_count = 0
_orig = m._fetch_url_content

def mock_fetch(url):
    global call_count
    call_count += 1
    if call_count > 10:
        raise RuntimeError(f"Stopped at call {call_count}")
    next_url = f"http://169.254.169.254/step{call_count}/"
    return f"""<html><head>
<meta http-equiv="refresh" content="0; url={next_url}">
</head><body>not a feed</body></html>""".encode()

m._fetch_url_content = mock_fetch

try:
    m.parse("http://attacker.com/loop")
except RuntimeError as e:
    print(f"CONFIRMED infinite loop: {e}")
finally:
    m._fetch_url_content = _orig
    print(f"Total fetches before stop: {call_count}")

# Output:
# CONFIRMED infinite loop: Stopped at call 11
# Total fetches before stop: 11

Each recursive call performs a real HTTP request (30 s timeout), HTML parsing, and a Python stack frame allocation. With Python's default recursion limit of 1000 and a 30 s per-request timeout, a single attacker request can hold a server thread busy for up to ~8 hours before a RecursionError is raised.

SSRF chain variant: The first response can be legitimate HTML redirecting to an internal address (http://192.168.1.1/), letting the redirect loop also serve as an SSRF bypass for targets that would otherwise be blocked by application-level URL validation applied only to the initial URL.

Impact

This is a denial-of-service vulnerability with a secondary SSRF-chaining impact. Any application that accepts user-supplied feed URLs and calls fastfeedparser.parse() is affected — including RSS aggregators, feed preview services, and "subscribe by URL" features. An attacker with no authentication can:

  • Hold a server worker thread indefinitely (one request per attacker connection)
  • Crash the worker process via RecursionError after ~1000 redirects
  • Use the redirect chain to pivot SSRF requests to internal network targets

Basic information

Type
reviewed
Severity
high
Advisory on GitHub
Open advisory ↗
Repository advisory
Open repository advisory ↗
Source code
Browse source ↗
Published (advisory)
2026-04-08 00:12:26 UTC
Updated
2026-06-06 00:24:06 UTC
GitHub reviewed
2026-04-08 00:12:26 UTC
NVD published
2026-04-07

EPSS Score

Score Percentile
0.08% 22.91%

CVSS Scores

Base score Version Severity Vector
7.5 3.1
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H Click to expand
Attack vector (AV:N)
Could be attacked over the internet or any normal routed network—not just someone sitting at the machine.
Attack complexity (AC:L)
Once they can reach the bug, pulling it off is straightforward—no weird race conditions or rare setup.
Privileges required (PR:N)
No account or special rights needed—anonymous or random user is enough.
User interaction (UI:N)
Nobody has to click “OK” or open a trap file; it can work without a victim helping.
Scope (S:U)
Damage stays in the same “trust bubble” as the broken component—no big spill into unrelated systems.
Confidentiality (C:N)
Doesn’t really leak secrets in a meaningful way.
Integrity (I:N)
Data isn’t meaningfully altered or forged.
Availability (A:H)
Could take the service down hard or make it unusable for people who depend on it.

Identifiers

CWEs

CWE id Name
CWE-400 Uncontrolled Resource Consumption
CWE-674 Uncontrolled Recursion

Credits

  • redyank (reporter)

Affected packages (1)

Vulnerable version ranges and first patched releases as published by GitHub.

Ecosystem Package Vulnerable range First patched Vulnerable functions
pip fastfeedparser <= 0.5.9 0.5.10

References

cvelogic Threat Intelligence