how
Pass
Audited by Gen Agent Trust Hub on Sep 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill instructions and metadata do not contain any malicious patterns, obfuscation, or unauthorized access attempts. The logic is strictly focused on technical analysis and explanation.
- [INDIRECT_PROMPT_INJECTION]: The skill processes external project files and reference documentation, creating an ingestion surface for untrusted data. This is common for analysis tools and is considered safe given the skill's limited capabilities.
- Ingestion points: Project source code (Luau modules, Instances), mapping files, and external references like
../../references/roblox-engineering.md. - Boundary markers: None explicitly defined, though it instructs the agent to cite exact paths and symbols for grounding.
- Capability inventory: Limited to technical tracing and 'scoped non-playtest edits' within the Studio environment; no network exfiltration or arbitrary shell execution capabilities are requested.
- Sanitization: Not present.
Audit Metadata