The Agent Skills Directory

[SAFE]: The skill content is a legitimate technical guide for reverse engineering and malware analysis. It does not include any malicious scripts, unauthorized data access commands, or hidden functionality. All tool references (e.g., Ghidra, IDA, Scylla) and API mentions are consistent with the stated purpose of analyzing binary files.
[PROMPT_INJECTION]: The skill contains no instructions designed to override agent safety protocols, bypass filters, or influence core behavioral logic. All instructions are focused on the task of binary analysis.
[DATA_EXFILTRATION]: There are no network-capable commands or instructions to access sensitive files like credentials, keys, or personal data. The workflow focuses on local binary processing and sandboxed execution.
[REMOTE_CODE_EXECUTION]: No external download or remote execution patterns were found. The skill references local tools for performing standard unpacking tasks.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted binary data, which is a potential ingestion point for indirect instructions. However, the workflow emphasizes the use of sandboxes and specialized analysis tools, and there is no evidence of vulnerable prompt interpolation or missing sanitization that would allow external content to influence the agent's control flow.
[DYNAMIC_CONTEXT_INJECTION]: No use of dynamic shell command execution syntax (e.g., !command) within the markdown was detected.

re-anti-analysis