The Agent Skills Directory

[COMMAND_EXECUTION]: The skill makes extensive use of the subprocess.run function across multiple files (generate.py, generate_image.py, generate_video.py, prompt_optimizer.py) to execute the system's pass utility and coordinate between local Python scripts.
[DATA_EXFILTRATION]: The tool retrieves API credentials from the system's pass password manager (pass api/xingjiabiapi) and transmits them in the Authorization header to the external domain xingjiabiapi.com. While this is functional for the skill's purpose, it involves automated extraction and external transmission of system-stored secrets.
[EXTERNAL_DOWNLOADS]: The skill downloads content from external, non-whitelisted domains including xingjiabiapi.com and s3.ffire.cc to save images and videos locally.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input by interpolating it into prompts sent to LLM optimizers and generative models, creating a vulnerability surface for indirect prompt injection.
Ingestion points: The user-provided prompt is ingested via sys.argv[1] in generate.py.
Boundary markers: None are present to delimit user input from system instructions.
Capability inventory: The skill has network access (requests), file system write access (open), and command execution capabilities (subprocess.run).
Sanitization: Relies on LLM-based instructions in prompt_optimizer.py to filter or rephrase content rather than deterministic sanitization methods.

multimodal-gen