The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses curl to perform a POST request to a target's xmlrpc.php endpoint. This command is intended to retrieve a list of supported XML-RPC methods for reconnaissance purposes and uses a placeholder URL (https://www.example.test).
[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection as it processes data from external targets.
Ingestion points: The skill downloads and saves XML data from a remote WordPress server to a local file (xmlrpc-methods.xml).
Boundary markers: No specific delimiters or "ignore embedded instructions" warnings are provided for the agent when it reads or processes the fetched XML content.
Capability inventory: The skill possesses the capability to create directories (mkdir) and perform network operations (curl).
Sanitization: There is no mechanism described for sanitizing or validating the XML content received from the remote target before it is saved or potentially processed by the agent.
[SAFE]: Despite the inclusion of red-teaming techniques, the skill is categorized as safe because it is explicitly framed as an authorized auditing tool. It includes multiple safeguards, such as requiring explicit permission for each step, mandating the use of synthetic/test identities, and strictly prohibiting the deployment of functional command shells in favor of benign markers.

wordpress-cors-xmlrpc-rce-chain