The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process user-provided academic content such as "claims, results, figures, notes, or Chinese drafts" as defined in the intake section of SKILL.md. While this establishes a data ingestion surface, the skill lacks explicit instructions for the agent to use boundary markers or to ignore embedded instructions within that untrusted data. Since the skill does not define access to dangerous tools (like shell execution or network writing), the risk is limited to manipulating the agent's text output.
[SAFE]: The skill references official educational resources for research writing on Notion and GitHub (found in README.md). These references are documented neutrally as sources for the writing strategies and do not involve suspicious network operations or exfiltration attempts.

nature-writing