onboard
Pass
Audited by Gen Agent Trust Hub on Jul 10, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: No malicious code, obfuscation, or unauthorized network activity was detected. The skill uses restricted tools (Read, Glob, Grep, Write) as defined in its frontmatter.- [PROMPT_INJECTION]: The skill ingests untrusted project data from various directories and interpolates it into an onboarding document without using boundary markers, creating a surface for indirect prompt injection if external project files contain malicious instructions.
- Ingestion points: Reads project overview from AGENTS.md and source files from src/, design/, and tests/ (SKILL.md).
- Boundary markers: No delimiters or 'ignore' instructions are used when summarizing external content.
- Capability inventory: Uses the Write tool to create new files in the production/onboarding/ directory.
- Sanitization: No input sanitization or validation is performed on the content read from project files.
Audit Metadata