The Agent Skills Directory

Indirect Prompt Injection (LOW): The skill possesses a surface for indirect prompt injection as it processes untrusted external data from user-provided papers and journal websites.
Ingestion points: The skill uses Read to ingest paper content and WebFetch to retrieve information from external journal websites.
Boundary markers: No specific delimiters or instructions to ignore embedded commands are present in the prompt structure.
Capability inventory: The skill utilizes WebSearch, WebFetch, and file system Read/Grep/Glob tools.
Sanitization: There is no explicit sanitization or filtering of the ingested content before it is processed by the LLM.
External Downloads (SAFE): While the skill uses WebFetch to gather journal data, it does not download executable code or packages; it only retrieves information for analysis purposes.
Data Exposure & Exfiltration (SAFE): The skill accesses local files via Read to analyze manuscripts, but there are no instructions or tools configured to exfiltrate this data to unauthorized third-party servers.

journal-target