The Agent Skills Directory

[DYNAMIC_EXECUTION]: The script retention.py dynamically modifies the Python module search path using sys.path.insert to load the deadair module from a computed relative directory (../yt-edit). This allows the execution of code from adjacent directories.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data, creating an ingestion surface. Ingestion points: Data is read from retention.csv and transcript.srt files. Boundary markers: The script output lacks explicit boundary markers or instructions to the agent to disregard instructions found within the data snippets. Capability inventory: The script is restricted to reading provided files and printing to standard output, without network access or file system write permissions. Sanitization: Subtitle content is truncated to 140 characters before being displayed, which limits the potential payload size for embedded instructions.

yt-retention