The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill recommends that users run npx workos@latest install to set up dependencies. This command fetches the WorkOS package from the npm registry. This is a vendor-owned resource and is safe in the context of the skill's purpose.
[REMOTE_CODE_EXECUTION]: The use of npx workos@latest install involves the execution of code downloaded from a remote registry. The skill correctly directs this action to the user for manual confirmation.
[COMMAND_EXECUTION]: A bundled utility script is executed via node references/scripts/query-spec.cjs to parse the OpenAPI specification for specific widget details. This is a local execution of a script provided with the skill.
[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection because it reads and processes data from external repository files to perform stack detection.
Ingestion points: Dependency manifests such as package.json, Gemfile, and pyproject.toml.
Boundary markers: No specific delimiters or safety warnings are used when the agent reads these files.
Capability inventory: The agent can execute shell commands (node, npx) and write code.
Sanitization: There is no evidence of sanitization for the data read from project files.

workos-widgets