The Agent Skills Directory

[SAFE]: The skill provides tools for analyzing ransomware binaries and encrypted artifacts. The analysis logic is entirely local and does not contain any network-capable code or exfiltration patterns.
[REMOTE_CODE_EXECUTION]: No remote code execution patterns were found. The skill relies on standard Python libraries (pefile, pycryptodome) for static analysis tasks.
[DATA_EXPOSURE_AND_EXFILTRATION]: The scripts (agent.py, process.py) read local files specified by the user to perform entropy calculations and pattern matching. There is no evidence of sensitive data being sent to external domains or hardcoded credentials.
[INDIRECT_PROMPT_INJECTION]: The skill provides the surface for processing untrusted data (malware samples). However, the processing is limited to regex matching, entropy calculation, and constant scanning. It lacks dangerous capabilities such as eval(), exec(), or subprocess spawning that would allow a malicious sample to gain control over the agent.
[DYNAMIC_CONTEXT_INJECTION]: The skill's frontmatter and instructions do not utilize any dynamic shell execution patterns.

reverse-engineering-ransomware-encryption-routine