The Agent Skills Directory

[PROMPT_INJECTION]: The skill is subject to indirect prompt injection because it fetches untrusted data from the external Zotero Web API for agent processing.\n
Ingestion points: Bibliographic metadata such as titles, abstracts, and full-text content are ingested via methods like zot.items(), zot.top(), and zot.fulltext_item() as documented in references/read-api.md and references/full-text.md.\n
Boundary markers: The skill does not implement or document the use of specific delimiters to separate untrusted API content from agent instructions.\n
Capability inventory: The agent has access to Bash, Write, and Edit tools (defined in SKILL.md), which could be misused if instructions hidden in Zotero items are followed.\n
Sanitization: There is no mention of sanitizing or validating the content retrieved from the Zotero API before it is presented to the agent.\n- [COMMAND_EXECUTION]: The documentation provides standard shell commands for installing the required Python packages and using the pyzotero command-line interface.\n
Evidence: uv add pyzotero and uv add "pyzotero[cli]" in SKILL.md and references/cli.md.\n- [EXTERNAL_DOWNLOADS]: The skill installs and uses standard libraries and CLI tools from the public PyPI registry.\n
Evidence: Downloads the pyzotero library as seen in the installation instructions.

alterlab-pyzotero