The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill downloads the @intelligentinternet/ii-commons package from the NPM registry and references the Intelligent-Internet/II-Commons-Skills GitHub repository.
[REMOTE_CODE_EXECUTION]: The skill utilizes npx to fetch and execute the @intelligentinternet/ii-commons CLI tool at runtime.
[COMMAND_EXECUTION]: The skill uses shell commands and spawnSync to execute search and retrieval tasks. It includes a security best practice by explicitly instructing the agent to use argument arrays rather than string interpolation to prevent command injection from user input.
[DATA_EXFILTRATION]: The skill connects to the external domain commons.ii.inc to perform research queries and download document metadata and Markdown content.
[PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it ingests untrusted data from academic (arXiv, PubMed) and policy corpora.
Ingestion points: External documents and metadata fetched via the ii-commons CLI (SKILL.md).
Boundary markers: The skill does not explicitly define delimiters for wrapping the retrieved research content in the provided examples.
Capability inventory: The skill uses spawnSync for command execution and has network access to commons.ii.inc.
Sanitization: The skill recommends using spawnSync with argument arrays to prevent injection into the command line, but does not specify sanitization for the content returned by the search tool.

ii-commons