The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill modernizes user-provided C++ code, which serves as an ingestion point for untrusted data. However, it lacks capabilities for automated execution, network communication, or file writing, functioning strictly as a guidance tool for the user. Findings indicate no exploitable logic for indirect injection.
[DATA_EXPOSURE_AND_EXFILTRATION]: No hardcoded credentials or sensitive file paths were detected. All network references target reputable sources including the official Trail of Bits blog, Google's security blog, OpenSSF hardening guides, and well-known developer tools like Godbolt and LLVM documentation. These are documented for informative purposes only.
[UNVERIFIABLE_DEPENDENCIES_AND_REMOTE_CODE_EXECUTION]: The skill provides examples of shell commands for compilation (g++, clang++) and binary verification (checksec, readelf). These are standard development workflows for software hardening and do not involve the automated download or execution of untrusted scripts. All referenced GitHub repositories belong to established organizations or the vendor itself.

modern-cpp