The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses subprocess.run() in scripts/publish_article.py to execute its own auxiliary Python scripts for parsing Markdown files and interacting with the system clipboard. These executions are limited to internal skill components.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied Markdown files and instructs the agent to analyze the content to generate article titles and image prompts. This creates a surface where embedded instructions in the Markdown could potentially influence the agent's actions.
Ingestion points: Markdown files processed by scripts/parse_markdown.py.
Boundary markers: None implemented for the text extraction process.
Capability inventory: Subprocess execution, browser automation via Patchright/Playwright, and file system access for image resolution.
Sanitization: Content is transformed into HTML for browser pasting but is not sanitized specifically for agent-facing prompts.
[EXTERNAL_DOWNLOADS]: The documentation specifies the installation of third-party libraries including Pillow, pyobjc-framework-Cocoa, and patchright to support image processing, macOS clipboard access, and browser automation.
[DYNAMIC_EXECUTION]: The automation scripts utilize page.evaluate() to run JavaScript code within the browser context, which is necessary for interacting with the rich text editor of the target website.

x-article-publisher