The Agent Skills Directory

[SAFE]: The skill serves exclusively as a documentation resource for Bun APIs. All code snippets are instructional examples for developers and do not contain malicious payloads or hidden logic.
[COMMAND_EXECUTION]: The documentation describes the Bun.$ shell API and Bun.spawn for process execution. It includes security best practices, specifically warning against shell injection and explaining how Bun's template literals provide automatic escaping for variables.
[EXTERNAL_DOWNLOADS]: The skill provides examples of fetching data from remote sources using fetch or curl for data processing pipelines. These examples target well-known or placeholder domains (e.g., example.com) and are standard patterns for the described functionality.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: The documentation covers APIs that read from Bun.stdin, local files via Bun.file(), and network sources via fetch.
Boundary markers: The documentation highlights that Bun.$ automatically escapes interpolated variables to prevent injection.
Capability inventory: The skill documents capabilities for shell execution (Bun.$, Bun.spawn), file writing (Bun.write), and database operations (bun:sqlite).
Sanitization: The reference materials explicitly warn about the dangers of using $.raw() and recommend using standard template interpolation for safety.

bun-api