analyzing-macro-malware-in-office-documents

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). This skill explicitly extracts and prints/deobfuscates VBA source and requires including deobfuscated code and decoded commands/URLs in the report, so any secrets embedded in macros would be output verbatim by the LLM (high exfiltration risk).

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.85). The required workflow runs scripts/agent.py on a user-supplied Office document and ingests outsider-authored free text from that document’s embedded VBA/OLE streams and OOXML XML/.rels contents (e.g., extract_vba_macros() reads vba_code, and check_dde()/check_external_relationships() decode XML text), which can then be included in the agent’s generated report/outputs and thus into any LLM context that consumes those outputs.