xref: /curl/docs/VULN-DISCLOSURE-POLICY.md (revision 86d33001)
1<!--
2Copyright (C) Daniel Stenberg, <daniel@haxx.se>, et al.
3
4SPDX-License-Identifier: curl
5-->
6
7# curl vulnerability disclosure policy
8
9This document describes how security vulnerabilities are handled in the curl
10project.
11
12## Publishing Information
13
14All known and public curl or libcurl related vulnerabilities are listed on
15[the curl website security page](https://curl.se/docs/security.html).
16
17Security vulnerabilities **should not** be entered in the project's public bug
18tracker.
19
20## Vulnerability Handling
21
22The typical process for handling a new security vulnerability is as follows.
23
24No information should be made public about a vulnerability until it is
25formally announced at the end of this process. That means, for example, that a
26bug tracker entry must NOT be created to track the issue since that makes the
27issue public and it should not be discussed on any of the project's public
28mailing lists. Messages associated with any commits should not make any
29reference to the security nature of the commit if done prior to the public
30announcement.
31
32- The person discovering the issue, the reporter, reports the vulnerability on
33  [HackerOne](https://hackerone.com/curl). Issues filed there reach a handful
34  of selected and trusted people.
35
36- Messages that do not relate to the reporting or managing of an undisclosed
37  security vulnerability in curl or libcurl are ignored and no further action
38  is required.
39
40- A person in the security team responds to the original report to acknowledge
41  that a human has seen the report.
42
43- The security team investigates the report and either rejects it or accepts
44  it. See below for examples of problems that are not considered
45  vulnerabilities.
46
47- If the report is rejected, the team writes to the reporter to explain why.
48
49- If the report is accepted, the team writes to the reporter to let them
50  know it is accepted and that they are working on a fix.
51
52- The security team discusses the problem, works out a fix, considers the
53  impact of the problem and suggests a release schedule. This discussion
54  should involve the reporter as much as possible.
55
56- The release of the information should be "as soon as possible" and is most
57  often synchronized with an upcoming release that contains the fix. If the
58  reporter, or anyone else involved, thinks the next planned release is too
59  far away, then a separate earlier release should be considered.
60
61- Write a security advisory draft about the problem that explains what the
62  problem is, its impact, which versions it affects, solutions or workarounds,
63  when the release is out and make sure to credit all contributors properly.
64  Figure out the CWE (Common Weakness Enumeration) number for the flaw. See
65  [SECURITY-ADVISORY](https://curl.se/dev/advisory.html) for help on creating
66  the advisory.
67
68- Request a CVE Id for the issue. curl is a CNA (CVE Numbering Authority) and
69  can request its own numbers.
70
71- Update the "security advisory" with the CVE number.
72
73- The security team commits the fix in a private branch. The commit message
74  should ideally contain the CVE number. If the severity level of the issue is
75  set to Low or Medium, the fix is allowed to get merged into the master
76  repository via a normal PR - but without mentioning it being a security
77  vulnerability.
78
79- The monetary reward part of the bug-bounty is managed by the Internet Bug
80  Bounty team and the reporter is asked to request the reward from them after
81  the issue has been completely handled and published by curl.
82
83- No more than 10 days before release, inform
84  [distros@openwall](https://oss-security.openwall.org/wiki/mailing-lists/distros)
85  to prepare them about the upcoming public security vulnerability
86  announcement - attach the advisory draft for information with CVE and
87  current patch. 'distros' does not accept an embargo longer than 14 days and
88  they do not care for Windows-specific flaws.
89
90- No more than 48 hours before the release, the private branch is merged into
91  the master branch and pushed. Once pushed, the information is accessible to
92  the public and the actual release should follow suit immediately afterwards.
93  The time between the push and the release is used for final tests and
94  reviews.
95
96- The project team creates a release that includes the fix.
97
98- The project team announces the release and the vulnerability to the world in
99  the same manner we always announce releases. It gets sent to the
100  curl-announce, curl-library and curl-users mailing lists.
101
102- The security webpage on the website should get the new vulnerability
103  mentioned.
104
105## security (at curl dot se)
106
107This is a private mailing list for discussions on and about curl security
108issues.
109
110Who is on this list? There are a couple of criteria you must meet, and then we
111might ask you to join the list or you can ask to join it. It really is not a
112formal process. We basically only require that you have a long-term presence
113in the curl project and you have shown an understanding for the project and
114its way of working. You must have been around for a good while and you should
115have no plans of vanishing in the near future.
116
117We do not make the list of participants public mostly because it tends to vary
118somewhat over time and a list somewhere only risks getting outdated.
119
120## Publishing Security Advisories
121
1221. Write up the security advisory, using markdown syntax. Use the same
123   subtitles as last time to maintain consistency.
124
1252. Name the advisory file after the allocated CVE id.
126
1273. Add a line on the top of the array in `curl-www/docs/vuln.pm`.
128
1294. Put the new advisory markdown file in the `curl-www/docs/` directory. Add it
130   to the git repository.
131
1325. Run `make` in your local web checkout and verify that things look fine.
133
1346. On security advisory release day, push the changes on the curl-www
135   repository's remote master branch.
136
137## HackerOne
138
139Request the issue to be disclosed. If there are sensitive details present in
140the report and discussion, those should be redacted from the disclosure. The
141default policy is to disclose as much as possible as soon as the vulnerability
142has been published.
143
144## Bug Bounty
145
146See [BUG-BOUNTY](https://curl.se/docs/bugbounty.html) for details on the
147bug bounty program.
148
149# Severity levels
150
151The curl project's security team rates security problems using four severity
152levels depending how serious we consider the problem to be. We use **Low**,
153**Medium**, **High** and **Critical**. We refrain from using numerical scoring
154of vulnerabilities.
155
156When deciding severity level on a particular issue, we take all the factors
157into account: attack vector, attack complexity, required privileges, necessary
158build configuration, protocols involved, platform specifics and also what
159effects a possible exploit or trigger of the issue can lead do, including
160confidentiality, integrity or availability problems.
161
162## Low
163
164This is a security problem that is truly hard or unlikely to exploit or
165trigger. Due to timing, platform requirements or the fact that options or
166protocols involved are rare etc. [Past
167example](https://curl.se/docs/CVE-2022-43552.html)
168
169## Medium
170
171This is a security problem that is less hard than **Low** to exploit or
172trigger. Less strict timing, wider platforms availability or involving more
173widely used options or protocols. A problem that usually needs something else
174to also happen to become serious. [Past
175example](https://curl.se/docs/CVE-2022-32206.html)
176
177## High
178
179This issue in itself a serious problem with real world impact. Flaws that can
180easily compromise the confidentiality, integrity or availability of resources.
181Exploiting or triggering this problem is not hard. [Past
182example](https://curl.se/docs/CVE-2019-3822.html)
183
184## Critical
185
186Easily exploitable by a remote unauthenticated attacker and lead to system
187compromise (arbitrary code execution) without requiring user interaction, with
188a common configuration on a popular platform. This issue has few restrictions
189and requirements and can be exploited easily using most curl configurations.
190[Past example](https://curl.se/docs/CVE-2000-0973.html)
191
192# Not security issues
193
194This is an incomplete list of issues that are not considered vulnerabilities.
195
196## Small memory leaks
197
198We do not consider a small memory leak a security problem; even if the amount
199of allocated memory grows by a small amount every now and then. Long-living
200applications and services already need to have counter-measures and deal with
201growing memory usage, be it leaks or just increased use. A small memory or
202resource leak is then expected to *not* cause a security problem.
203
204Of course there can be a discussion if a leak is small or not. A large leak
205can be considered a security problem due to the DOS risk. If leaked memory
206contains sensitive data it might also qualify as a security problem.
207
208## Never-ending transfers
209
210We do not consider flaws that cause a transfer to never end to be a security
211problem. There are already several benign and likely reasons for transfers to
212stall and never end, so applications that cannot deal with never-ending
213transfers already need to have counter-measures established.
214
215If the problem avoids the regular counter-measures when it causes a never-
216ending transfer, it might be a security problem.
217
218## Not practically possible
219
220If the flaw or vulnerability cannot practically get executed on existing
221hardware it is not a security problem.
222
223## API misuse
224
225If a reported issue only triggers by an application using the API in a way
226that is not documented to work or even documented to not work, it is probably
227not going to be considered a security problem. We only guarantee secure and
228proper functionality when the APIs are used as expected and documented.
229
230There can be a discussion about what the documentation actually means and how
231to interpret the text, which might end up with us still agreeing that it is a
232security problem.
233
234## Local attackers already present
235
236When an issue can only be attacked or misused by an attacker present on the
237local system or network, the bar is raised. If a local user wrongfully has
238elevated rights on your system enough to attack curl, they can probably
239already do much worse harm and the problem is not really in curl.
240
241## Experiments
242
243Vulnerabilities in features which are off by default (in the build) and
244documented as experimental, are not eligible for a reward and we do not
245consider them security problems.
246
247## URL inconsistencies
248
249URL parser inconsistencies between browsers and curl are expected and are not
250considered security vulnerabilities. The WHATWG URL Specification and RFC
2513986+ (the plus meaning that it is an extended version) [are not completely
252interoperable](https://github.com/bagder/docs/blob/master/URL-interop.md).
253
254Obvious parser bugs can still be vulnerabilities of course.
255
256## Visible command line arguments
257
258The curl command blanks the contents of a number of command line arguments to
259prevent them from appearing in process listings. It does not blank all
260arguments even if some of them that are not blanked might contain sensitive
261data. We consider this functionality a best-effort and omissions are not
262security vulnerabilities.
263
264 - not all systems allow the arguments to be blanked in the first place
265 - since curl blanks the argument itself they area readable for a short moment
266   no matter what
267 - virtually every argument can contain sensitive data, depending on use
268 - blanking all arguments would make it impractical for users to differentiate
269   curl command lines in process listings
270
271## Busy-loops
272
273Busy-loops that consume 100% CPU time but eventually end (perhaps due to a set
274timeout value or otherwise) are not considered security problems. Applications
275are supposed to already handle situations when the transfer loop legitimately
276consumes 100% CPU time, so while a prolonged such busy-loop is a nasty bug, we
277do not consider it a security problem.
278
279## Saving files
280
281curl cannot protect against attacks where an attacker has write access to the
282same directory where curl is directed to save files.
283
284## Tricking a user to run a command line
285
286A creative, misleading or funny looking command line is not a security
287problem. The curl command line tool takes options and URLs on the command line
288and if an attacker can trick the user to run a specifically crafted curl
289command line, all bets are off. Such an attacker can just as well have the
290user run a much worse command that can do something fatal (like
291`sudo rm -rf /`).
292
293## Terminal output and escape sequences
294
295Content that is transferred from a server and gets displayed in a terminal by
296curl may contain escape sequences or use other tricks to fool the user. This
297is curl working as designed and is not a curl security problem. Escape
298sequences, moving cursor, changing color etc, is also frequently used for
299good. To reduce the risk of getting fooled, save files and browse them after
300download using a display method that minimizes risks.
301