The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external content which creates a surface for indirect prompt injection attacks.
Ingestion points: The skill ingests data from external web pages via gooseworks:data_call_provider, media transcripts through gooseworks:media_analyze, and competitor data.
Boundary markers: It utilizes a tagging system ([PROOF NEEDED], [CUSTOMER APPROVAL NEEDED]) to identify and isolate unverified claims or external data within drafts.
Capability inventory: The agent can write to the workspace via gooseworks:file_write and update the brand's knowledge base using gooseworks:knowledge_import.
Sanitization: The skill implements a strict human-in-the-loop policy ("Nothing is published — you edit and post") and a multi-pass substance and readability review to filter out malicious or unintended instructions.
[EXTERNAL_DOWNLOADS]: The skill performs research by fetching content from public websites and social media platforms (LinkedIn, Threads, X) using authorized vendor tools. These operations are intended for gathering market insights and competitor information.

make-content