The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes external system utilities to perform document validation and auditing functions.
ooxml/scripts/pack.py: Executes soffice (LibreOffice) in headless mode to convert repacked documents into HTML as a verification step to ensure the output file is not corrupt.
ooxml/scripts/validation/redlining.py: Executes git diff with word-level precision to validate that document modifications are properly recorded as tracked changes.
[PROMPT_INJECTION]: The skill architecture handles untrusted external data (Word documents), establishing an attack surface for indirect prompt injection that is mitigated through technical isolation.
Ingestion points: ooxml/scripts/unpack.py and scripts/document.py read and extract the XML structure from user-provided Office files.
Boundary markers: The skill instructs the agent to treat document content strictly as XML elements, separating data from instructional logic.
Capability inventory: The skill has the capability to write files to the workspace and execute document conversion/diffing tools via subprocess.
Sanitization: All XML parsing is initiated using the defusedxml library, which protects against XML External Entity (XXE) and billion laughs attacks.
[SAFE]: No malicious patterns, data exfiltration, or obfuscation were detected. The skill focuses on standard technical document processing and employs appropriate security measures for its intended use case.

docx