The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses curl and python3 to execute searches and process structured data from academic databases. It includes instructions for running specific shell pipelines to filter and format research metadata.
[EXTERNAL_DOWNLOADS]: The skill performs network operations to fetch metadata from established academic services, including arxiv.org, semanticscholar.org, and dblp.org (documented in rules/literature-search.md).
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface through the processing of untrusted external data:
Ingestion points: Research paper abstracts and literature metadata are fetched directly from external APIs and integrated into the agent's context (e.g., in rules/literature-search.md).
Boundary markers: There are no explicit delimiters or instructions provided to the agent to treat fetched literature content as untrusted or to ignore instructions embedded within research abstracts.
Capability inventory: The skill has access to bash command execution (curl, python3) and file system operations as described in its core workflow.
Sanitization: The skill does not implement automated sanitization or filtering for the external text retrieved during the literature search process.

academic-writer