bun-file-io
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOWNO_CODE
Full Analysis
- [Prompt Injection] (SAFE): No instructions found attempting to override agent behavior or bypass safety filters.
- [Data Exposure & Exfiltration] (SAFE): While the skill discusses file reading and writing, it does not include hardcoded credentials or unauthorized network exfiltration patterns.
- [Remote Code Execution] (SAFE): Mention of Bun.spawn is for legitimate subprocess management as per documentation; no remote script execution or piped downloads are present.
- [Obfuscation] (SAFE): No encoded or hidden content detected.
- [Indirect Prompt Injection] (INFO): The skill describes patterns for ingesting external data and executing commands. 1. Ingestion points: Bun.file(path).text() and Bun.readableStreamToText(stream) in SKILL.md. 2. Boundary markers: None mentioned. 3. Capability inventory: Bun.write, Bun.spawn, and Bun.file().delete() in SKILL.md. 4. Sanitization: Not mentioned in instructions. As this is a documentation skill, it lacks an active implementation to exploit.
Audit Metadata