The Agent Skills Directory

[DATA_EXFILTRATION]: The skill provides instructions for targeting sensitive files during exploitation, including system files like /etc/passwd, /etc/hostname, and /proc/self/environ.
[DATA_EXFILTRATION]: Instructs the agent on how to retrieve cloud and local credential files such as ~/.aws/credentials and ~/.ssh/id_rsa.
[COMMAND_EXECUTION]: Provides a Python one-liner for a local HTTP redirect server used in Time-of-Check-Time-of-Use (TOCTOU) exploitation.
[EXTERNAL_DOWNLOADS]: References external security testing and Out-of-Band (OOB) services including interactsh.com, oastify.com, r3dir.me, and 1u.ms.
[PROMPT_INJECTION]: The skill creates an indirect prompt injection surface by processing untrusted data from target applications and external state management systems.
Ingestion points: Target server responses, input parameters (URL, src, href, etc.), and the get_state_summary() MCP server.
Boundary markers: No explicit boundary markers or instructions to disregard embedded commands in the processed data are present.
Capability inventory: Execution of security tools (burpsuite, ssrfmap, gopherus, interactsh) and writing data to the engagement/evidence/ directory.
Sanitization: No sanitization or validation of external content is specified before processing or logging.

ssrf