The Agent Skills Directory

[COMMAND_EXECUTION]: The skill makes extensive use of the Bash tool to execute local utility scripts such as enum_imports.py and to run inline Python code blocks for binary analysis and pattern searching.
[REMOTE_CODE_EXECUTION]: The instructions direct the agent to dynamically generate Python exploit scripts in the ./exploits/ directory and execute them using Bash to verify suspected vulnerabilities. This creates a high-capability execution loop based on analyzed data.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection due to its processing of untrusted data from debuggee binaries. Ingestion points: Reads PE headers, import/export names, and memory contents via enum_imports.py and x64dbg tools. Boundary markers: None identified in instructions to separate data from instructions. Capability inventory: Significant capabilities including shell execution via Bash and memory modification via debugger tools. Sanitization: No validation or sanitization is performed on symbols or strings retrieved from the target binary.
[EXTERNAL_DOWNLOADS]: The enum_imports.py script suggests the installation of the lief library via pip if it is missing from the environment. This is a well-known security library but involves an external package installation.

vuln-hunter