Repository navigation
Conversation
load_resources reads each wordlist combination once per process and deduplicates it, so the larger list doesn't multiply per-event I/O or set rebuilds. Sourced from SecLists xato-net top-1M; superset of the old top-100k.
Cracks NextAuth v4 (A256GCM) and Auth.js v5 (A256CBC-HS512) session cookies via HKDF-derived keys against known/leaked secrets. Adds JWE helpers and Supabase/Directus default JWT secrets.
Bumps [django](https://github.com/django/django) from 5.2.15 to 5.2.16. - [Commits](django/django@5.2.15...5.2.16) --- updated-dependencies: - dependency-name: django dependency-version: 5.2.16 dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <support@github.com>
…jango-5.2.16 build(deps): bump django from 5.2.15 to 5.2.16
- add __COMPRESSED_VSTATE to the compressed viewstate carve regex and YARA prefilter; it was in neither, so pages using that field name were never detected - _carve_body walks re.finditer instead of taking only re.search's first match, so an empty decoy __VIEWSTATE no longer shadows a real payload - reshape the aspnet_compressedviewstate and jsf_viewstate carve regexes to assert the field name in a lookahead: allows attributes between name= and value=, and fixes quadratic backtracking
…state-carve Fix carve misses on __COMPRESSED_VSTATE fields
Add NextAuth/Auth.js JWE session token module
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #432 +/- ##
==========================================
- Coverage 99.45% 99.29% -0.16%
==========================================
Files 30 31 +1
Lines 3102 3259 +157
==========================================
+ Hits 3085 3236 +151
- Misses 17 23 +6 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
Bumps [cryptography](https://github.com/pyca/cryptography) from 48.0.1 to 50.0.0. - [Changelog](https://github.com/pyca/cryptography/blob/main/CHANGELOG.rst) - [Commits](pyca/cryptography@48.0.1...50.0.0) --- updated-dependencies: - dependency-name: cryptography dependency-version: 50.0.0 dependency-type: indirect ... Signed-off-by: dependabot[bot] <support@github.com>
Bumps [django](https://github.com/django/django) from 5.2.15 to 5.2.16. - [Commits](django/django@5.2.15...5.2.16) --- updated-dependencies: - dependency-name: django dependency-version: 5.2.16 dependency-type: direct:production ... Signed-off-by: dependabot[bot] <support@github.com>
Two defects made the positional CLI path silently mishandle valid input, so a viewstate that cracked via --url would report nothing when the same viewstate and generator were passed by hand: - check_secret_args was 3, so the 4th positional was sliced away by check_all_modules. check_secret() accepts (viewstate, generator, url, ViewStateUserKey), but url and userkey could never be supplied together -- the userkey vanished with no error. Raised to 4. - the generator pattern matched uppercase hex only. A lowercase generator fell through to the ViewStateUserKey branch, leaving the generator at its "0000" default and checking the wrong thing, then reporting "No secrets found". Matching is now case-insensitive and resolve_args canonicalizes to uppercase so downstream path brute-forcing and result strings agree. Also document the remaining asymmetry, which is a real constraint rather than a bug: DOTNET45 derives the validation key from the page path via the SP800-108 KDF purposes, and DOTNET40 with IsolateApps mixes in an app-path hash, so neither can be validated without the URL. This is the usual reason --url succeeds where passing the viewstate and generator alone does not.
Bumps [sqlparse](https://github.com/andialbrecht/sqlparse) from 0.5.5 to 0.6.0. - [Changelog](https://github.com/andialbrecht/sqlparse/blob/master/CHANGELOG) - [Commits](andialbrecht/sqlparse@0.5.5...0.6.0) --- updated-dependencies: - dependency-name: sqlparse dependency-version: 0.6.0 dependency-type: indirect ... Signed-off-by: dependabot[bot] <support@github.com>
…rse-0.6.0 build(deps): bump sqlparse from 0.5.5 to 0.6.0
…o-5.2.16 build(deps): bump django from 5.2.15 to 5.2.16
…ography-50.0.0 build(deps): bump cryptography from 48.0.1 to 50.0.0
Modules that brute-force top_250000_passwords.txt pay it on every check_secret() call that gets past identify(). For most that is a cheap HMAC per candidate, but the JSF viewstate 3DES path derives a SHA1PRNG key per candidate in pure Python, costing ~13s per call locally and far more on CI under coverage's tracer. The 3.10 job timed out at 40 minutes on all three attempts as a result. Add a trim_wordlist fixture and use it in the affected test files to cut the list down to the entries each test actually needs. The shipped wordlists and module behaviour are unchanged.
Bumps [urllib3](https://github.com/urllib3/urllib3) from 2.7.0 to 2.8.0. - [Release notes](https://github.com/urllib3/urllib3/releases) - [Changelog](https://github.com/urllib3/urllib3/blob/main/CHANGES.rst) - [Commits](urllib3/urllib3@2.7.0...2.8.0) --- updated-dependencies: - dependency-name: urllib3 dependency-version: 2.8.0 dependency-type: indirect ... Signed-off-by: dependabot[bot] <support@github.com>
The autouse `_trim_wordlists` fixture was pasted into seven test files, with
only nextauth_test.py varying the call. Replace the copies with a single
autouse fixture in conftest.py that reads a registered `trim_wordlist`
marker, so opting a file in is one line:
pytestmark = pytest.mark.trim_wordlist
and a file that has to crack a secret deep in the list passes it as a marker
arg, as nextauth_test.py does with "secret". The marker also accepts
`wordlist=` and `context=`, which pass through to the existing fixture.
The fixture resolves `trim_wordlist` lazily via getfixturevalue, so unmarked
tests never build it and keep the shared resource cache rather than paying a
cache clear on every teardown.
twos_complement() took the signed value of a byte via bin() -> string replace -> int(str, 2) -> to_bytes -> from_bytes, and updateState() called it twice per byte. Deriving one 24-byte key ran it 60 times, so a full sweep of top_250000_passwords.txt made 20M calls and updateState accounted for 23s of a 40s profile. Replace it with a 256-entry lookup table of the same values. Output is byte-identical -- verified exhaustively over all 256 byte values and against 45 pre-change key vectors -- and updateState drops to 2.4s, with the JSF 3DES sweep going from 13.3s to 9.6s. The remaining cost is pycryptodome's DES3 key setup, not our code. Adds tests pinning the derivation.
Speed up tests by trimming brute-force wordlists
Bumps [django](https://github.com/django/django) from 5.2.16 to 5.2.17. - [Commits](django/django@5.2.16...5.2.17) --- updated-dependencies: - dependency-name: django dependency-version: 5.2.17 dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <support@github.com>
Bumps blasthttp from 0.9.0 to 0.10.0. --- updated-dependencies: - dependency-name: blasthttp dependency-version: 0.10.0 dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <support@github.com>
Bumps [django](https://github.com/django/django) from 5.2.15 to 5.2.17. - [Commits](django/django@5.2.15...5.2.17) --- updated-dependencies: - dependency-name: django dependency-version: 5.2.17 dependency-type: direct:production ... Signed-off-by: dependabot[bot] <support@github.com>
…lasthttp-0.10.0 build(deps): bump blasthttp from 0.9.0 to 0.10.0
…b3-2.8.0 build(deps): bump urllib3 from 2.7.0 to 2.8.0
…jango-5.2.17 build(deps): bump django from 5.2.16 to 5.2.17
Bumps [pytest-mock](https://github.com/pytest-dev/pytest-mock) from 3.15.1 to 3.16.0. - [Release notes](https://github.com/pytest-dev/pytest-mock/releases) - [Changelog](https://github.com/pytest-dev/pytest-mock/blob/main/CHANGELOG.rst) - [Commits](pytest-dev/pytest-mock@v3.15.1...v3.16.0) --- updated-dependencies: - dependency-name: pytest-mock dependency-version: 3.16.0 dependency-type: direct:development update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <support@github.com>
…ytest-mock-3.16.0 build(deps-dev): bump pytest-mock from 3.15.1 to 3.16.0
…o-5.2.17 build(deps): bump django from 5.2.15 to 5.2.17
Speed up Java SHA1PRNG key derivation
fix ASP.NET viewstate CLI argument handling
Two related defects in the JWT module, both rooted in handing pyjwt a token other than the one that was found. XMLDSIG: tokens that spell "alg" as a URI are signed over that header, but swap_algorithm rewrote it to the JWS short name before verification. That changes the signing input, so a genuine URI-algorithm token could never be cracked, and the hashcat command we printed was guaranteed to fail. Teach a private pyjwt instance the URI names and verify the original bytes instead. The instance is private so the extra names cannot affect anything else in the process that imports pyjwt. The existing xmldsig fixture is an HS256 token with its header replaced and the original signature kept, so it only verifies against the rewritten form. A URI-spelled header can legitimately be a rewrite of a token signed under the short name, so that form is still tried, after the token as issued. Non-canonical signatures: base64url leaves unused trailing bits, so several spellings of a signature decode to the same bytes. pyjwt 2.14 rejects all but the canonical one (CVE-2026-102269), which made it refuse tokens we can still crack, and took the header down with it even though the header itself was fine. Parse the header directly, and keep a canonically re-encoded copy for pyjwt alone. Only the signature segment is touched, so the decoded signature and the signing input are both unchanged. Findings and hashcat commands now quote the product exactly as it was found; the rewritten forms never leave the verification path. Verified against pyjwt 2.13.0 and 2.15.0.
Bumps [pycryptodome](https://github.com/Legrandin/pycryptodome) from 3.23.0 to 3.24.0. - [Release notes](https://github.com/Legrandin/pycryptodome/releases) - [Changelog](https://github.com/Legrandin/pycryptodome/blob/master/Changelog.rst) - [Commits](Legrandin/pycryptodome@v3.23.0...v3.24.0) --- updated-dependencies: - dependency-name: pycryptodome dependency-version: 3.24.0 dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <support@github.com>
Bumps [django](https://github.com/django/django) from 5.2.17 to 5.2.18. - [Commits](django/django@5.2.17...5.2.18) --- updated-dependencies: - dependency-name: django dependency-version: 5.2.18 dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <support@github.com>
…ut-and-noncanonical-sigs Verify JWTs as issued instead of rewriting them
…ycryptodome-3.24.0 build(deps): bump pycryptodome from 3.23.0 to 3.24.0
…jango-5.2.18 build(deps): bump django from 5.2.17 to 5.2.18
Bumps [werkzeug](https://github.com/pallets/werkzeug) from 3.1.8 to 3.1.9. - [Release notes](https://github.com/pallets/werkzeug/releases) - [Changelog](https://github.com/pallets/werkzeug/blob/main/CHANGES.rst) - [Commits](pallets/werkzeug@3.1.8...3.1.9) --- updated-dependencies: - dependency-name: werkzeug dependency-version: 3.1.9 dependency-type: indirect ... Signed-off-by: dependabot[bot] <support@github.com>
…eug-3.1.9 build(deps): bump werkzeug from 3.1.8 to 3.1.9
en0f
approved these changes
Oct 14, 2026
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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.
Release 1.3.0.
New module
NextAuthpassive module: checks NextAuth.js v4 (A256GCM) and Auth.js v5 (A256CBC-HS512)dirJWE session cookies for a known or weakNEXTAUTH_SECRET/AUTH_SECRET, via HKDF-derived keys. Adds JWE helpers and Supabase/Directus default JWT secrets (Add NextAuth/Auth.js JWE session token module #431).Fixes
__COMPRESSED_VSTATEfields — the field name was in neither the carve regex nor the YARA prefilter, so pages using it were never detected. Also fixes first-match shadowing, where an empty decoy__VIEWSTATEhid a real payload, and quadratic backtracking in the compressed-viewstate and JSF carve regexes (Fix carve misses on __COMPRESSED_VSTATE fields #430).--urlreported nothing when the same viewstate and generator were passed positionally.check_secret_argswas 3, so the 4th positional (ViewStateUserKey) was silently sliced away; and the generator pattern matched uppercase hex only, so a lowercase generator fell through to the wrong branch and left the generator at its0000default (fix ASP.NET viewstate CLI argument handling #436).Resources and performance
load_resourcesso the larger list doesn't multiply per-event I/O or set rebuilds. Re-points eight passive modules to the new list. Sourced from SecLists xato-net top-1M; superset of the old top-100k.twos_complement()'s bin/str/int round-trip with a 256-entry lookup table. Output is byte-identical, verified exhaustively over all 256 byte values and against 45 pre-change key vectors.updateStatedrops from 23s to 2.4s and the JSF 3DES sweep from 13.3s to 9.6s (Speed up Java SHA1PRNG key derivation #440).trim_wordlistpytest marker. The 3.10 CI job had been timing out at 40 minutes on all three attempts. Shipped wordlists and module behaviour are unchanged (Speed up tests by trimming brute-force wordlists #439).Dependencies
Release
__version__to 1.3.0.Merging this to
maintriggers thepublishjob intests.yaml: it runs after the test matrix, compares__version__against the latest tag (currently1.2.1), then builds, uploads to PyPI, and pushes a1.3.0tag.