The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or unauthorized data exfiltration attempts were detected. The skill's functionality is consistent with its stated purpose as a research management tool.
[PROMPT_INJECTION]: The skill processes untrusted external data from PDFs and Zotero metadata during search and RAG operations (e.g., zot pdf, zot workspace query, zot ask). While this presents a surface for indirect prompt injection, it is inherent to the tool's function. The skill includes instructions for the agent to synthesize answers from citation-keyed evidence, which aids in maintaining context boundaries.
[COMMAND_EXECUTION]: The skill uses the zot CLI for data operations. It includes safety features like --dry-run and --idempotency-key for write operations. The zot bridge install command generates a plugin for Zotero desktop but requires manual user intervention for the final installation step in the Zotero GUI, serving as a security checkpoint.
[EXTERNAL_DOWNLOADS]: The skill interacts with the official Zotero Web API and can fetch academic PDFs via Zotero's desktop resolvers or paper metadata via the Semantic Scholar API. These network operations are limited to well-known academic services necessary for the skill's operation.

zotero-cli-cc