The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the installation of the sciverse-mcp-server package from the NPM registry and facilitates communication with the sciverse.space API.
[COMMAND_EXECUTION]: The skill's setup instructions involve shell commands for global package installation (npm install -g) and executing the MCP server using npx.
[DATA_EXFILTRATION]: The skill uses an API token (SCIVERSE_API_TOKEN) to authenticate requests to the service's official backend at sciverse.space.
[PROMPT_INJECTION]: The skill ingests external content from academic papers, which presents a surface for indirect prompt injection where instructions embedded in papers could attempt to influence the agent.
Ingestion points: Abstracts and full-text fragments retrieved through the semantic_search and read_content tools (SKILL.md).
Boundary markers: Absent. The instructions do not provide specific delimiters or ignore-instructions warnings for the retrieved academic data.
Capability inventory: The agent can read content, perform searches, and process scientific data through its standard reasoning cycle.
Sanitization: Absent. The skill relies on the inherent formatting of academic papers and does not specify content filtering.

sciverse