The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses npx to download and execute the @jfrog/agent-guard package from the vendor's official registry at https://releases.jfrog.io/artifactory/api/npm/coding-agents-npm/. This is used for catalog discovery and repository provisioning.
[COMMAND_EXECUTION]: The skill extensively uses the jf CLI for operations like install, update, publish, and list. It also interacts with harness-specific CLI tools (e.g., claude, cursor) to manage plugin states within those environments.
[INDIRECT_PROMPT_INJECTION]: The skill processes and displays data fetched from external sources in the JFrog AI Catalog.
Ingestion points: Data is ingested via npx @jfrog/agent-guard --list-skills and jf skills list commands described in SKILL.md and reference files.
Boundary markers: Results are presented to the user using structured Markdown tables.
Capability inventory: The skill has permissions to write to the filesystem and execute commands via the installation of new skills and plugins (references/installing-skills.md, references/installing-plugins.md).
Sanitization: The skill parses output from the CLI tools but does not explicitly implement additional sanitization or escaping for the strings returned from the catalog descriptions before rendering.

jfrog-ai-catalog