The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the agent to install the mcp-zotero package (with rag and ocr extras) using the uv tool. This package originates from a source that is not included in the pre-defined list of trusted organizations or well-known services.
[COMMAND_EXECUTION]: The agent is directed to execute shell commands to verify existing installations (uv tool list | grep mcp-zotero) and to install or upgrade software components.
[CREDENTIALS_UNSAFE]: The setup workflow involves requesting the user's Zotero Library ID and local API key (ZOTERO_LOCAL_KEY). These sensitive credentials are subsequently written to a local .mcp.json file. While intended for local operation, the handling of secrets through an automated agent setup process requires user caution.
[PROMPT_INJECTION]: The skill facilitates Retrieval-Augmented Generation (RAG) by indexing PDF documents. This creates a surface for indirect prompt injection, where malicious instructions embedded in documents could influence the agent's behavior during search tasks.
Ingestion points: PDF files retrieved from the user's Zotero storage directory (typically ~/Zotero/storage).
Boundary markers: The instructions do not define specific delimiters or instructions to ignore embedded content when processing text chunks from search results.
Capability inventory: The system includes tools for indexing libraries, performing semantic searches, expanding text context, and finding similar document chunks.
Sanitization: There is no evidence of sanitization, filtering, or validation of the text extracted from PDFs before it is presented to the agent's context.

zotero-rag