The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill retrieves and processes full-text scientific articles and PDFs from external sources via Europe PMC. This untrusted data is ingested into the agent's context, creating a surface for potential indirect prompt injection where adversarial text in a paper could attempt to override agent instructions.
Ingestion points: scripts/europepmc_api.py downloads external content via the get_fulltext and download_pdf commands.
Boundary markers: None explicitly mentioned in the processing script to distinguish paper content from instructions.
Capability inventory: The skill performs network requests and writes data to the local filesystem.
Sanitization: The script performs XML tag stripping but does not sanitize the text content for potential prompt injection patterns.
[EXTERNAL_DOWNLOADS]: The skill communicates with official Europe PMC and EMBL-EBI domains (europepmc.org and ebi.ac.uk) to fetch scientific metadata and literature.
[COMMAND_EXECUTION]: The skill uses uv run to execute the local utility scripts/europepmc_api.py which facilitates all interactions with the Europe PMC API.

literature-search-europepmc