The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates the analysis of untrusted local binaries (e.g., using strings, grep, and debuggers like gdb). This creates a surface for indirect prompt injection, where an attacker could embed malicious instructions within a binary's strings to influence the agent's behavior during analysis.
Ingestion points: Local binary files (e.g., sample.exe) whose contents are read into the agent's context via tools like strings and Ghidra.
Boundary markers: The skill does not prescribe the use of delimiters or specific instructions to the agent to ignore content found within the analyzed binaries.
Capability inventory: The environment includes shell access and the ability to execute system tools and debuggers.
Sanitization: There are no instructions for sanitizing or filtering the output of string extraction tools before the agent processes them.

re-license