[DRAFT] feat: begin modernization of symmetric crypto [#268][#289][#357] #411
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Add pytest fixture parameterization to test under both dev & production settings.
Suggested in review: scidsg#411 (comment) scidsg#411 (comment) scidsg#411 (comment) Co-authored-by: Jeremy Moore <jeremymoore@gmail.com>
|
| GitGuardian id | GitGuardian status | Secret | Commit | Filename | |
|---|---|---|---|---|---|
| - | - | Generic High Entropy Secret | 7f35ab2 | docker-compose.yaml | View secret |
| - | - | Generic High Entropy Secret | d6af3e3 | dev_env.sh | View secret |
| - | - | Generic High Entropy Secret | 0924906 | dev_env.sh | View secret |
| - | - | Generic High Entropy Secret | 7f35ab2 | dev_env.sh | View secret |
| - | - | Generic High Entropy Secret | 294338b | dev_env.sh | View secret |
| - | - | Generic High Entropy Secret | 294338b | docker-compose.yaml | View secret |
🛠 Guidelines to remediate hardcoded secrets
- Understand the implications of revoking this secret by investigating where it is used in your code.
- Replace and store your secrets safely. Learn here the best practices.
- Revoke and rotate these secrets.
- If possible, rewrite git history. Rewriting git history is not a trivial act. You might completely break other contributing developers' workflow and you risk accidentally deleting legitimate data.
To avoid such incidents in the future consider
- following these best practices for managing and storing secrets including API keys and other credentials
- install secret detection on pre-commit to catch secret before it leaves your machine and ease remediation.
🦉 GitGuardian detects secrets in your source code to help developers and security teams secure the modern development process. You are seeing this because you or someone else with access to this repository has authorized GitGuardian to scan your pull request.
rmlibre
commented
Jul 18, 2024
There was a problem hiding this comment.
How does having a single shared salt compare to generating a per-user salt? Does using a kdf which incorporates the domain (and possible aad) obviate this need?
Although since we may still want to generate a per-user salt for passwords to combat precompute attacks.
rmlibre
commented
Jul 22, 2024
rmlibre
added a commit
to rmlibre/hushline
that referenced
this pull request
Jul 22, 2024
Cherry-picked from (b2cb88e) on PR branch [scidsg#411]
3 tasks
rmlibre
added a commit
to rmlibre/hushline
that referenced
this pull request
Jul 23, 2024
Cherry-picked from (e173071) on PR branch [scidsg#411]
Add a hushline specific domain label, & add name of destination hash to pre-hash domain label. These are additional good practice measures to mitigate hash shucking by eliminating external datasets from being possible sources of correlation. References: a. https://www.youtube.com/watch?v=OQD3qDYMyYQ b. https://security.stackexchange.com/a/234795 Reference (b) excerpt: "But the crucial point is that [a pre-hash] can be an otherwise uncracked [pre-hash] that happens to be present in some other leak, which you can then attack at massively increased speeds."
…scidsg#411]" This reverts commit e173071.
…idsg#411] The embedded timestamps in Fernet ciphertexts can be utilized in the case of server compromise to correlate analyzed web traffic originating from a targeted user, even when using Tor, with the encrypted database values hosted by servers running hushline. This commit switches from the Fernet cipher to the AEAD cipher ChaCha20Poly1305 which doesn't automatically generate timestamps, removing this source of unintended activity tracking correlation. Using the DKNA method (derived inputs: key, nonce, associated data), the 12 byte nonce is easily upgraded to a 32 byte salt, making repeats infeasible: >=2**-85.33 chance for each additional message after ~2**85.33 encryptions in the same context. This is referred to as the optimal bound for birthday bound security. The DKNA method also provides resistance to commitment attacks by binding the non-message inputs together into pseudo-random internal cipher states. This switch also comes with space efficiency improvements for the database. 32.65% more plaintext can fit into the same fields, due to the new raw bytes ciphertext being on average ~31.12% more efficient than Fernet's base64 encoded ciphertext. Fernet: ------- (Base64 encoded) 1 B | 8 B | 16 B | X B | 32 B Version ‖ Timestamp ‖ IV ‖ Ciphertext ‖ HMAC ChaCha20Poly1305 with DKNA: ------- (Raw bytes) ------- 32 B | X B | 16 B Salt ‖ Ciphertext ‖ Poly-Tag References: https://dataprot.net/articles/no-log-vpn/ https://martinolivier.com/open/timestamps.pdf https://github.com/Attacks-on-Tor/Attacks-on-Tor#correlation-attacks https://github.com/rmlibre/hushline/blob/7799379088e833282a0a8b9e8f9fd21756321522/docs/2-threat-model.md?plain=1#L75-L79 https://soatok.blog/2024/07/01/blowing-out-the-candles-on-the-birthday-bound/ https://crypto.stackexchange.com/q/112497
rmlibre
added a commit
to rmlibre/hushline
that referenced
this pull request
Jul 30, 2024
Quick-fix addressing privacy concerns discussed in PR branch [scidsg#411] commit (c772523). References: scidsg@c772523 Co-authored-by: Jeremy Moore <jeremymoore@gmail.com>
2 tasks
This was referenced Sep 29, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
Generally, this PR is aimed at bringing best practices to the handling of secrets. That is done through making them more inaccessible, or by clearing them from memory as soon as possible. This PR also brings domain separation, input canonicalization, context commitment, & authenticated associated data to distinct encryption / decryption contexts.
More details & updates to come. Comments & reviews are open.
Domain Separation:
When using cryptography to authenticate data, or derive secrets, etc, it's important to tie those processes to their intended usage contexts by declaring the context with an unambiguous label. This allows both the production & consumption of cryptographic artifacts to ensure that they're in agreement over the label (domain) the artifact is supposed to be used within.
References:
What is meant by domain separation ...
What non-trivial benefit does including ...
Separate Your Domains: NIST PQC ...
Input Canonicalization:
Cryptography works best when its users are 100% sure & clear about what information they're processing. That can be subtlety, and surprisingly, challenging to do.
Example:
Say we're not using canonical input encoding, and we're hashing (H) a username: "green" & an email address: "housekeeper[at]email[dot]com". Here, if the delineation is unclear, then the information is not clear. Another user with username: "greenhouse" and email address: "keeper[at]email[dot]com" will produce the same hash.
References:
What is Canonicalization Attack?
Canonicalization Attacks Against ...
AEAD & Context Commitment:
Confusingly, the two most widely deployed authenticated ciphers (AES-GCM & ChaCha20-Poly1305) do not create ciphertexts which commit to a key. Which means that knowing the key allows one to create a single ciphertext which successfully decrypts to various different plaintext messages. In most cases this isn't considered an issue because it requires knowledge of the key. But, it does violate intuitive notions of what it means for something to be authentic when forgeries are easy to produce. It has also been shown to lead to practical problems:
Because Fernet uses HMAC-SHA-256, instead of the universal hashes of the previously discussed ciphers, it doesn't suffer from this problem. But, it also doesn't have a direct interface for authenticating domain labels or other associated data. Borrowing partially from Efficient Schemes for Committing ..., which proposes deriving a fresh key by running a keyed PRF over all of the inputs, it can elegantly support the authentication of additional values.
References:
Key Committing AEADs
How to Abuse and Fix Authenticated Encryption ...
Succinctly-Committing Authenticated Encryption
Key commitment in GCM (or AEAD in general)
Remediations
... detailed descriptions coming soon
Passing Workflows: