The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external binary files (referenced as 'sample.exe'), which serves as an ingestion point for untrusted data that could potentially contain malicious payloads.
Ingestion points: The agent is instructed to load, analyze, and modify external executable samples provided by the user or found in the environment (SKILL.md, Operation Steps 1-3).
Boundary markers: The instructions explicitly mandate the use of isolated environments ("[[re-sandbox]]") and the creation of file backups ("cp sample.exe sample.bak") to mitigate operational risks.
Capability inventory: The skill utilizes high-privilege capabilities including file system writing via rizin -w and process debugging via gdb or x64dbg.
Sanitization: There is no sanitization performed on the binary data itself, as the content is the direct object of the analysis.
[COMMAND_EXECUTION]: The skill provides commands for installing and executing various system-level tools for binary analysis and modification.
Evidence: Instructions include package manager commands (e.g., apt install hexedit, brew install rizin) and direct shell execution for patching (e.g., rizin -w sample.exe followed by wx commands to overwrite bytes).

re-patching