no-vibe-code
Pass
Audited by Gen Agent Trust Hub on Sep 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from local source files during the linting process, creating a surface for indirect prompt injection via the snippets returned to the agent.
- Ingestion points: slop-check.mjs reads content from files in target directories.
- Boundary markers: Absent. The findings are reported directly to the agent context without delimiters or warnings to ignore instructions within snippets.
- Capability inventory: Limited to local file reads and console logging within slop-check.mjs.
- Sanitization: The tool extracts raw substrings as snippets without escaping or validating the content for instructional patterns.
- [REMOTE_CODE_EXECUTION]: No remote code execution or external dependency risks were identified. The included Node.js script uses only built-in modules and operates on local files.
- [DATA_EXFILTRATION]: No network access or data exfiltration logic was detected in the script or instructions.
Audit Metadata