The Agent Skills Directory

[DATA_EXFILTRATION]: The skill is configured to log granular data including internal system file paths (Source_File_Path), cell coordinates, and raw financial values to external log files (CSV, JSONL, and Excel). While this is the intended primary purpose of the skill for audit traceability, it results in the exposure of internal directory structures and sensitive financial information in the output logs.
[INDIRECT_PROMPT_INJECTION]: The skill acts as a data processing agent that ingests content from external sources to generate audit trails.
Ingestion points: Data is read from external files and system outputs as specified in the Source_File_Path and Source_Value_Read parameters.
Boundary markers: The instructions do not explicitly provide delimiters or specific instructions for the agent to disregard instructions that may be embedded within the spreadsheet data being audited.
Capability inventory: The skill maintains the ability to write to multiple file formats (Excel, CSV, JSONL) on the local file system.
Sanitization: While the skill includes a specific protocol to redact credentials (e.g., passwords, PINs), it lacks explicit sanitization or escaping mechanisms for general data read from source files before it is committed to the audit logs.

financial-audit-trail-guardian