The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on executing the sherlock command-line tool in the terminal to retrieve account information. This introduces a risk if the user-supplied username is not properly sanitized, potentially leading to command injection despite the use of double quotes in the template.
[EXTERNAL_DOWNLOADS]: The skill provides instructions for installing the sherlock-project tool using package managers like pip or pipx, or by pulling a Docker image from sherlock/sherlock.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from a large number of external websites and user input, which could be used to influence the agent's behavior if malicious content is returned or provided.
Ingestion points: User input (usernames) and responses from 400+ external social media and web platforms (SKILL.md).
Boundary markers: The procedure uses double quotes ("<username>") in the command construction to delimit user input, but it lacks explicit instructions to sanitize the input for shell metacharacters.
Capability inventory: Terminal execution capability via the terminal tool (SKILL.md).
Sanitization: No specific sanitization or filtering is defined for the content retrieved from external sites or for the username provided by the user.

sherlock