The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute core tools like hashcat and john with the dangerouslyDisableSandbox: true flag, which bypasses the execution environment's security controls.
[COMMAND_EXECUTION]: The skill requires the use of sudo to perform administrative tasks such as installing the john, wordlists, and seclists packages.
[COMMAND_EXECUTION]: There is a potential for command injection as the skill constructs shell commands (e.g., zip2john protected.zip > hash.txt) using user-supplied file paths and filenames without explicit sanitization or escaping.
[COMMAND_EXECUTION]: Risk of indirect prompt injection through data ingestion.
Ingestion points: The skill processes potentially untrusted data from password hashes and encrypted archives provided by the user or other skills.
Boundary markers: None identified; the skill lacks delimiters or specific instructions to ignore embedded commands within the data being processed.
Capability inventory: The skill utilizes the Bash tool with broad permissions (dangerouslyDisableSandbox: true) for tool execution and file manipulation.
Sanitization: Absent; the skill does not mention validation or escaping of the ingested hash data or archive filenames before passing them to the shell.
[EXTERNAL_DOWNLOADS]: Fetches software source code from the official Openwall repository on GitHub for building the John the Ripper tool.

credential-recovery