The Agent Skills Directory

[PROMPT_INJECTION]: The skill uses an "AI LOAD INSTRUCTION" block to establish a specific expert persona and provide instructions for processing hash attack scenarios.- [UNVERIFIABLE_DEPENDENCIES_AND_REMOTE_CODE_EXECUTION]: The documentation provides shell command examples for external cryptanalysis tools such as HashPump and fastcoll, and the Python code references the third-party hlextend library.- [DATA_EXPOSURE_AND_EXFILTRATION]: Section 5.2 implements an HMAC timing attack script using the requests library, which enables automated network requests to target URLs.- [INDIRECT_PROMPT_INJECTION]: The skill contains a vulnerability surface in Section 5.2 where the hmac_timing_attack function ingests untrusted URL and data parameters without sanitization or boundary markers, combined with network access capabilities via requests.get.

hash-attack-techniques