The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill includes functionality to download media files (images and videos) from external URLs to a local session cache directory for processing. This is implemented in scripts/aigc_textgen.py via the download_media function using the urllib library.
[COMMAND_EXECUTION]: The skill utilizes Python scripts to interact with the LinkFox API. Instructions in SKILL.md guide the agent to execute these scripts and suggest using shell tools like jq to parse and chain outputs between different AI agent skills.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external sources, specifically through the prompt and imageUrls parameters. These inputs are passed to an LLM for generation and analysis, presenting a standard indirect prompt injection surface.
Ingestion points: prompt string and imageUrls array in scripts/aigc_textgen.py and SKILL.md.
Boundary markers: None explicitly used in the script level; the skill acts as a client to a remote AI service.
Capability inventory: Performs network requests (API calls and media downloads) and file system operations (caching results and media in a local linkfox/ directory).
Sanitization: Implements a custom line-break encoding/decoding mechanism (⏎ U+23CE) to ensure text content can be safely passed through shell variables without breaking JSON structures.

linkfox-aigc-textgen