The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides numerous shell commands for environment checking, interactive setup wizards, platform-specific logins, and diagnostic checks.
[EXTERNAL_DOWNLOADS]: Instructions are provided to download and install the 'feedgrab' Python package and its optional dependencies from a package registry.
[INDIRECT_PROMPT_INJECTION]: The skill configures a tool intended to ingest untrusted data from external platforms, creating a potential surface for indirect injection.
Ingestion points: The tool fetches content from Twitter/X, WeChat, Xiaohongshu, Feishu, Reddit, and other news or social media platforms.
Boundary markers: The instructions do not define specific delimiters or directives to the agent for handling potentially malicious instructions embedded in scraped content.
Capability inventory: The skill and associated tool have the capability to write to the local file system (output directory), perform network requests, and execute browser automation via Playwright/Patchright.
Sanitization: No explicit content sanitization, filtering, or validation steps are mentioned in the setup configuration for processed external data.

feedgrab-setup