memory-forensics
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data extracted from memory dumps, which may contain instructions designed to influence the agent's behavior.
- Ingestion points: The MemoryAnalyzer class loads memory images from local file paths and extracts strings, URLs, and network connection details.
- Boundary markers: No specific delimiters or safety instructions are provided to help the agent distinguish between data and embedded instructions within the memory strings.
- Capability inventory: The skill includes methods for writing local files, performing YARA scans, and querying external threat intelligence feeds.
- Sanitization: No evidence of sanitization or escaping of extracted strings is present before they are passed into the agent's context.
- [DATA_EXFILTRATION]: The skill identifies and extracts sensitive credentials, including NTLM hashes, Kerberos tickets, and LSA secrets from system memory.
- Evidence: The CredentialExtractor class provides automated methods for harvesting authentication artifacts. This is consistent with the skill's primary purpose and findings are exported to local files.
- [COMMAND_EXECUTION]: The MemoryAcquisition module generates shell commands for memory acquisition tools.
- Evidence: The generate_command method produces command strings for tools like winpmem and DumpIt to facilitate acquisition by the user or agent.
Audit Metadata