The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes Python scripts (scripts/aigc_textgen.py, scripts/onboarding.py) to interact with the LinkFox API. These scripts perform standard HTTP requests using urllib and requests. No arbitrary shell command execution from untrusted sources was detected.\n- [EXTERNAL_DOWNLOADS]: The download_media function in scripts/aigc_textgen.py allows downloading media files (images/videos) from URLs provided by the user. These downloads are scoped to the local session directory for processing and do not involve remote code execution. Neutral documentation of this feature is maintained as it is central to the skill's purpose.\n- [DATA_EXFILTRATION]: The skill requires an API key (LINKFOX_AGENT_API_KEY) to authorize requests to the vendor's gateway (tool-gateway.linkfox.com). This is standard authentication for a cloud-based service. There is no evidence of exfiltrating local sensitive files (like SSH keys or AWS credentials).\n- [PROMPT_INJECTION]: The skill is a conduit for LLM text generation. While it processes user prompts, it does not contain instructions designed to bypass the underlying model's safety filters or override system behavior beyond its specific text-generation task.\n- [ONBOARDING_SAFETY]: The scripts/onboarding.py script manages account-related tasks like sending SMS codes and querying orders. These operations are conducted against the vendor's official domains (api.linkfox.com, agent-api.linkfox.com). The script includes a render_qr function which depends on the qrcode and pillow libraries, both of which are well-known and standard for this purpose.

linkfox-aigc-textgen