The Agent Skills Directory

[COMMAND_EXECUTION]: The skill contains explicit instructions for the agent to execute local shell and Python commands, such as python run_pipeline.py and python calendar_generator.py, to manage the research workflow.
[DATA_EXFILTRATION]: The skill references and accesses specific local file paths, including a hardcoded absolute directory (/Users/shaileshsingh/cowriting system/research-engine) and internal data stores like latest_scrape.json and calendar.json. This may expose the local file system structure and sensitive research data to the agent.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data harvested from YouTube comments and scraped channel content. * Ingestion points: External data is ingested via files such as latest_scrape.json and latest_comments.json. * Boundary markers: No specific delimiters or instructions are provided to separate the ingested data from the agent's core instructions. * Capability inventory: The skill provides capabilities for shell command execution and local file access. * Sanitization: There is no evidence of sanitization or filtering of the scraped content before it is used for script generation, allowing potential malicious prompts in comments to influence agent behavior.

youtube-script-master