terms-of-use-writer

Pass

Audited by Gen Agent Trust Hub on Jun 25, 2026

Risk Level: SAFE
Full Analysis
  • [DATA_EXFILTRATION]: Broad Read Access. The skill is designed to inspect a wide range of project files, including database schemas (schema.prisma), API routes, server logic, and application configurations (package.json, next.config.*), to identify product functionality for legal drafting. This is an intended function for the skill's purpose.
  • [COMMAND_EXECUTION]: File System and Code Modification. The skill writes new documents (TERMS_OF_USE.md) and modifies application source code to implement legal pages (e.g., in app/, pages/, or src/ directories), following existing project patterns.
  • [PROMPT_INJECTION]: Indirect Prompt Injection Surface. The skill is susceptible to indirect prompt injection because it processes untrusted content from the codebase. 1. Ingestion points: Project source code, documentation, and configuration files (referenced in SKILL.md). 2. Boundary markers: Absent; the skill does not use specific delimiters to distinguish between data and instructions. 3. Capability inventory: File writing and source code modification (SKILL.md Step 4 and Step 5). 4. Sanitization: Absent; the skill does not filter or sanitize the content extracted from project files.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 25, 2026, 02:08 PM
Security Audit — agent-trust-hub — terms-of-use-writer