Skip to content

Commit 769c5cb

Browse files
jaimergpbeckermr
andcommitted
add blog/2021-09-24-travis-security.md
Co-authored-by: beckermr <[email protected]>
1 parent 3f9aec4 commit 769c5cb

File tree

1 file changed

+54
-0
lines changed

1 file changed

+54
-0
lines changed

blog/2021-09-24-travis-security.md

Lines changed: 54 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,54 @@
1+
---
2+
authors:
3+
- beckermr
4+
tags: [security]
5+
---
6+
# Travis CI Security Incident
7+
8+
On September 9, 2021 one of our core devs discovered that artifacts
9+
building on Travis CI were being uploaded to our conda channel from PRs
10+
running on forked repositories. A quick investigation revealed that
11+
Travis CI was passing encrypted secrets to PR builds on forks. Further
12+
examination of our logs and artifacts indicated that this had been
13+
happening since about September 3, 2021. This security bug was
14+
subsequently confirmed by Travis CI. See this
15+
[CVE](https://nvd.nist.gov/vuln/detail/CVE-2021-41077) for more details
16+
on this incident. **As far as we know, there were no actual exploits
17+
against conda-forge which used this vulnerability.**
18+
19+
<!--truncate-->
20+
21+
## Our Response
22+
23+
We took the following steps to respond to this incident.
24+
25+
1. We immediately turned off all builds on Travis CI by suspending the
26+
Travis CI GitHub App.
27+
2. We immediately disclosed the bug to Travis CI through our contacts
28+
there.
29+
3. Once Travis CI indicated to us that they were ready, we rotated all
30+
feedstock tokens and later our anaconda.org token for our staging
31+
channel. The anaconda.org token for the main `conda-forge` channel
32+
was never disclosed in this incident. Further, only ~70 feedstocks
33+
had their tokens exposed in this incident.
34+
4. We examined our artifacts and marked as broken any artifacts that
35+
were uploaded from PRs. We think we found everything, but we are not
36+
completely sure. Our criterion for marking things broken was more
37+
generous than it needed to be.
38+
5. We issued PRs to rebuild any broken artifacts via our bots.
39+
6. We put in changes to `conda-smithy` to help prevent inadvertent
40+
uploads of artifacts from PRs in the future.
41+
42+
## Closing Thoughts & What can you do?
43+
44+
I (MRB) want to recognize the quick work of our core dev team in
45+
handling this incident. It goes without saying that the public nature of
46+
`conda-forge`'s infrastructure carries risks. On the other hand, by
47+
being public, anyone can look and verify our artifact builds. Security
48+
for `conda-forge` is about reducing risk and we will continue to do our
49+
best.
50+
51+
Our best defense against security incidents in `conda-forge` is *you*!
52+
Our feedstock maintainers are in the best position to notice incidents
53+
and issues. Please responsibly report anything you find to us at
54+

0 commit comments

Comments
 (0)