skills/lobbi-docs/claude/graphql/Gen Agent Trust Hub

graphql

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [Indirect Prompt Injection] (LOW): Identified a vulnerability surface where untrusted data enters the agent context. (1) Ingestion points: ExecuteTaskInput message field in SKILL.md. (2) Boundary markers: Absent. (3) Capability inventory: Bash, Write, and Edit tools enabled in frontmatter. (4) Sanitization: Absent in the provided templates.
  • [SAFE] (SAFE): The provided code snippets for Python (Strawberry), Node.js (Apollo), and React are standard boilerplate for GraphQL development with no detected malicious patterns, obfuscation, or unauthorized network operations.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:25 PM