The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides instructions and automated routines for installing necessary CLI dependencies, such as the Grok CLI from x.ai and various Printing Press research tools from the author's own @mvanhorn npm scope. These are documented as official or vendor-owned resources.
[DATA_EXFILTRATION]: To provide zero-config authentication for services like X/Twitter and Truth Social, the skill extracts session cookies from local browser databases (Firefox, Chrome, Safari). This operation is a core feature disclosed in the documentation and via the preflight and doctor commands.
[COMMAND_EXECUTION]: The skill utilizes numerous external binaries for its retrieval pipeline, including yt-dlp for video transcripts, gh for GitHub data, and openssl for cookie decryption on macOS. These subprocesses are used for their intended purpose within the research workflow.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the web and social media and feeds it into LLMs for planning and reranking. It mitigates this risk by using <untrusted_content> boundary markers and explicit instructions to the model to treat the content as data rather than instructions.
[CREDENTIALS_UNSAFE]: A hardcoded public bearer token is present in the vendored bird-search Node.js library. This is a well-known public constant required for the scraper's operation and does not involve the exposure of private user secrets.

last30days