The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and parse external forensic artifacts, specifically raw binary $MFT files and CSV outputs from forensic tools. This creates an inherent attack surface where malicious data (e.g., specially crafted filenames or metadata within a disk image) could potentially be interpreted as instructions by an AI agent during analysis.
Ingestion points: scripts/agent.py reads raw binary MFT data; scripts/process.py reads CSV records.
Boundary markers: Not explicitly defined in the parsing scripts.
Capability inventory: Both scripts have the ability to write analysis results to the local file system using json.dump.
Sanitization: The scripts decode binary data (UTF-16LE) but do not perform secondary validation for prompt injection patterns before including them in the JSON reports.
[SAFE]: The Python scripts (scripts/agent.py and scripts/process.py) use standard library modules (struct, json, csv, argparse) to perform legitimate forensic parsing. No suspicious imports, obfuscation, or network operations were found.

analyzing-mft-for-deleted-file-recovery