The Agent Skills Directory

[DYNAMIC_EXECUTION]: The retention.py script dynamically modifies the Python module search path at runtime using sys.path.insert to include the relative directory ../yt-edit. It then imports a module named deadair from this computed path. While this is likely used for modularity within a local skill set, loading code from computed paths is a recognized security risk.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data provided via CSV and transcript (SRT) files, creating a potential surface for indirect prompt injection.
Ingestion points: Data is read from files passed as command-line arguments to retention.py (e.g., retention.csv, transcript.srt).
Boundary markers: The skill does not use specific delimiters or instructions to the agent to ignore potentially malicious content within these files.
Capability inventory: The script's capabilities are restricted; it performs data analysis and prints results to stdout. It does not perform network operations, file system writes, or subprocess execution.
Sanitization: The CSV parser attempts to convert inputs to floats, which serves as a basic validation. The transcript text is truncated to 140 characters, but no further sanitization or escaping is performed before output.

yt-retention