The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill provides extensive reference material and specific payloads for achieving remote code execution across multiple platforms. This includes OS command injection one-liners (reverse shells) in reference/os-command-injection-cheat-sheet.md, SSTI class traversal payloads for Python, Java, and Ruby in reference/ssti-advanced.md, and database-specific RCE techniques such as MSSQL's xp_cmdshell and SQLite's load_extension() in reference/sql-injection-advanced.md.
[COMMAND_EXECUTION]: Several functional Python scripts are included (e.g., in reference/nosql-injection-advanced.md and reference/sql-injection-advanced.md) that automate the process of testing for vulnerabilities. These scripts construct and execute complex network requests using the requests library to probe target systems with injection payloads.
[DATA_EXFILTRATION]: The skill documentation and scripts describe sophisticated methods for extracting sensitive data from compromised systems. This includes character-by-character extraction using regular expressions in NoSQL (reference/nosql-injection-advanced.md), and time-based or out-of-band (DNS/HTTP) exfiltration techniques for SQL and XXE vulnerabilities found throughout the reference documents.
[CREDENTIALS_UNSAFE]: The documentation identifies high-value targets for file retrieval, including sensitive system files like /etc/shadow, private SSH keys (id_rsa), and application environment files (.env), providing exact payloads to access these locations in the OS Command Injection and SQL Injection reference files.

injection