The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides instructions to install the pyzotero, bibtexparser, and python-dotenv packages. These are well-established, standard libraries in the Python ecosystem for API interaction, bibliographic parsing, and environment configuration.\n- [PROMPT_INJECTION]: The skill enables the agent to retrieve and process data from external Zotero libraries, including item metadata, notes, and full-text indexed PDF content (e.g., via zot.items(), zot.fulltext_item(), and CLI searches). This creates a surface for indirect prompt injection, as malicious instructions could be embedded in bibliographic entries or attachments which the agent might then execute.\n
Ingestion points: Data retrieved via read-api.md, full-text.md, and the CLI search functionality described in cli.md.\n
Boundary markers: The skill does not implement specific boundary markers or instructions to ignore embedded commands in the retrieved bibliographic data.\n
Capability inventory: The skill has access to Bash, Write, and Edit tools according to the YAML frontmatter.\n
Sanitization: There is no indication of content sanitization or validation performed on the external data before it enters the agent's context.

pyzotero