investigating-windows-endpoints
Pass
Audited by Gen Agent Trust Hub on Jul 31, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill mentions and provides links to several external forensic toolsets, including the Eric Zimmerman suite (PECmd, AmcacheParser, MFTECmd, etc.), KAPE, Chainsaw, Hayabusa, and Plaso. These are well-known, community-standard tools for digital forensics and incident response (DFIR) and are used here within their intended professional context.
- [PROMPT_INJECTION]: This skill presents a surface for indirect prompt injection because it processes untrusted forensic artifacts from potentially compromised systems. An adversary could plant malicious commands or instructions within log entries or registry keys to influence the agent's analysis.
- Ingestion points: Ingests untrusted data from .evtx logs, Registry hives, $MFT, and USN Journals using various parsers.
- Boundary markers: No specific delimiters or boundary markers are defined to isolate the ingested data from the agent's command context.
- Capability inventory: The skill utilizes command execution to run local forensic binaries and performs extensive read operations on sensitive system files.
- Sanitization: There are no instructions for sanitizing or validating the content of the artifacts before processing.
Audit Metadata