laravel-taste-validator
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided Laravel application code and interpolates it into 21 potential prompts dispatched to a sub-agent (
vibemastery-toolkit:laravel-architect). This creates a vulnerability surface where malicious instructions embedded in the code under review could attempt to influence the sub-agent's evaluation. - Ingestion points: User-provided code samples and files as described in the
SKILL.mdworkflow. - Boundary markers: The skill uses structural labels like
PRINCIPLE CRITERIA:andCODE TO REVIEW:to separate instructions from user data, but it lacks strict sanitization or instructions for the sub-agent to ignore embedded commands within the code. - Capability inventory: The primary skill does not possess sensitive capabilities such as network access, filesystem writes, or subprocess execution; its primary function is text processing and agent delegation.
- Sanitization: No sanitization or escaping of the input code is performed before it is passed to the sub-agent.
Audit Metadata