axiom-codable

skills-sh:charleswiltgen_axiom__axiom-codable

View source

0/100

First Seen

Feb 19, 2026

Last Scanned

Feb 19, 2026

Findings

Score

0/100

MEDIUM 19

Findings (19)

MEDIUM

Base64-encoded instructions PROMPT_INJECTION_009

prompt-injection L180

Detects instructions to decode and execute base64 content

decoder.keyDecodingStrategy = .convertFromSnakeCase

FIX

Remove text that simulates multi-turn conversations or fake user/assistant message pairs. These patterns attempt to manipulate the agent by fabricating conversation history.

FP?

Likely FP if the matched text is documentation showing example conversations or API request/response formats for illustration purposes.

MEDIUM

Base64-encoded instructions PROMPT_INJECTION_009

prompt-injection L301

Detects instructions to decode and execute base64 content

decoder.singleValueContainer

FIX

Remove text that simulates multi-turn conversations or fake user/assistant message pairs. These patterns attempt to manipulate the agent by fabricating conversation history.

FP?

Likely FP if the matched text is documentation showing example conversations or API request/response formats for illustration purposes.

MEDIUM

Base64-encoded instructions PROMPT_INJECTION_009

prompt-injection L327

Detects instructions to decode and execute base64 content

decoder.dateDecodingStrategy

FIX

Remove text that simulates multi-turn conversations or fake user/assistant message pairs. These patterns attempt to manipulate the agent by fabricating conversation history.

FP?

Likely FP if the matched text is documentation showing example conversations or API request/response formats for illustration purposes.

MEDIUM

Base64-encoded instructions PROMPT_INJECTION_009

prompt-injection L331

Detects instructions to decode and execute base64 content

decoder.dateDecodingStrategy

FIX

Remove text that simulates multi-turn conversations or fake user/assistant message pairs. These patterns attempt to manipulate the agent by fabricating conversation history.

FP?

Likely FP if the matched text is documentation showing example conversations or API request/response formats for illustration purposes.

MEDIUM

Base64-encoded instructions PROMPT_INJECTION_009

prompt-injection L335

Detects instructions to decode and execute base64 content

decoder.dateDecodingStrategy = .millisecondsSince1970

FIX

Remove text that simulates multi-turn conversations or fake user/assistant message pairs. These patterns attempt to manipulate the agent by fabricating conversation history.

FP?

Likely FP if the matched text is documentation showing example conversations or API request/response formats for illustration purposes.

MEDIUM

Base64-encoded instructions PROMPT_INJECTION_009

prompt-injection L343

Detects instructions to decode and execute base64 content

decoder.dateDecodingStrategy

FIX

Remove text that simulates multi-turn conversations or fake user/assistant message pairs. These patterns attempt to manipulate the agent by fabricating conversation history.

FP?

Likely FP if the matched text is documentation showing example conversations or API request/response formats for illustration purposes.

MEDIUM

Base64-encoded instructions PROMPT_INJECTION_009

prompt-injection L346

Detects instructions to decode and execute base64 content

decoder.dateDecodingStrategy

FIX

Remove text that simulates multi-turn conversations or fake user/assistant message pairs. These patterns attempt to manipulate the agent by fabricating conversation history.

FP?

Likely FP if the matched text is documentation showing example conversations or API request/response formats for illustration purposes.

MEDIUM

Base64-encoded instructions PROMPT_INJECTION_009

prompt-injection L347

Detects instructions to decode and execute base64 content

decoder.singleValueContainer

FIX

Remove text that simulates multi-turn conversations or fake user/assistant message pairs. These patterns attempt to manipulate the agent by fabricating conversation history.

FP?

Likely FP if the matched text is documentation showing example conversations or API request/response formats for illustration purposes.

MEDIUM

Base64-encoded instructions PROMPT_INJECTION_009

prompt-injection L380

Detects instructions to decode and execute base64 content

decoder.dateDecodingStrategy

FIX

Remove text that simulates multi-turn conversations or fake user/assistant message pairs. These patterns attempt to manipulate the agent by fabricating conversation history.

FP?

Likely FP if the matched text is documentation showing example conversations or API request/response formats for illustration purposes.

MEDIUM

Base64-encoded instructions PROMPT_INJECTION_009

prompt-injection L389

Detects instructions to decode and execute base64 content

decoder.dateDecodingStrategy

FIX

Remove text that simulates multi-turn conversations or fake user/assistant message pairs. These patterns attempt to manipulate the agent by fabricating conversation history.

FP?

Likely FP if the matched text is documentation showing example conversations or API request/response formats for illustration purposes.

MEDIUM

Base64-encoded instructions PROMPT_INJECTION_009

prompt-injection L414

Detects instructions to decode and execute base64 content

decoder.singleValueContainer

FIX

Remove text that simulates multi-turn conversations or fake user/assistant message pairs. These patterns attempt to manipulate the agent by fabricating conversation history.

FP?

Likely FP if the matched text is documentation showing example conversations or API request/response formats for illustration purposes.

MEDIUM

Base64-encoded instructions PROMPT_INJECTION_009

prompt-injection L462

Detects instructions to decode and execute base64 content

decoder.singleValueContainer

FIX

Remove text that simulates multi-turn conversations or fake user/assistant message pairs. These patterns attempt to manipulate the agent by fabricating conversation history.

FP?

Likely FP if the matched text is documentation showing example conversations or API request/response formats for illustration purposes.

MEDIUM

Base64-encoded instructions PROMPT_INJECTION_009

prompt-injection L525

Detects instructions to decode and execute base64 content

decoder.userInfo[configurationUserInfoKey

FIX

Remove text that simulates multi-turn conversations or fake user/assistant message pairs. These patterns attempt to manipulate the agent by fabricating conversation history.

FP?

Likely FP if the matched text is documentation showing example conversations or API request/response formats for illustration purposes.

MEDIUM

Base64-encoded instructions PROMPT_INJECTION_009

prompt-injection L530

Detects instructions to decode and execute base64 content

decode<T: DecodableWithConfiguration

FIX

Remove text that simulates multi-turn conversations or fake user/assistant message pairs. These patterns attempt to manipulate the agent by fabricating conversation history.

FP?

Likely FP if the matched text is documentation showing example conversations or API request/response formats for illustration purposes.

MEDIUM

Base64-encoded instructions PROMPT_INJECTION_009

prompt-injection L536

Detects instructions to decode and execute base64 content

decoder.userInfo[Self.configurationUserInfoKey

FIX

Remove text that simulates multi-turn conversations or fake user/assistant message pairs. These patterns attempt to manipulate the agent by fabricating conversation history.

FP?

Likely FP if the matched text is documentation showing example conversations or API request/response formats for illustration purposes.

MEDIUM

Base64-encoded instructions PROMPT_INJECTION_009

prompt-injection L537

Detects instructions to decode and execute base64 content

decoder.decode(ConfigurationDecodingWrapper

FIX

Remove text that simulates multi-turn conversations or fake user/assistant message pairs. These patterns attempt to manipulate the agent by fabricating conversation history.

FP?

Likely FP if the matched text is documentation showing example conversations or API request/response formats for illustration purposes.

MEDIUM

Base64-encoded instructions PROMPT_INJECTION_009

prompt-injection L708

Detects instructions to decode and execute base64 content

decoder.dateDecodingStrategy

FIX

Remove text that simulates multi-turn conversations or fake user/assistant message pairs. These patterns attempt to manipulate the agent by fabricating conversation history.

FP?

Likely FP if the matched text is documentation showing example conversations or API request/response formats for illustration purposes.

MEDIUM

Base64-encoded instructions PROMPT_INJECTION_009

prompt-injection L718

Detects instructions to decode and execute base64 content

decoder.dateDecodingStrategy

FIX

Remove text that simulates multi-turn conversations or fake user/assistant message pairs. These patterns attempt to manipulate the agent by fabricating conversation history.

FP?

Likely FP if the matched text is documentation showing example conversations or API request/response formats for illustration purposes.

MEDIUM

Base64-encoded instructions PROMPT_INJECTION_009

prompt-injection L719

Detects instructions to decode and execute base64 content

decoder.singleValueContainer

FIX

Remove text that simulates multi-turn conversations or fake user/assistant message pairs. These patterns attempt to manipulate the agent by fabricating conversation history.

FP?

Likely FP if the matched text is documentation showing example conversations or API request/response formats for illustration purposes.