security-and-hardening
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is an instructional resource that provides comprehensive guidance on implementing security controls for web applications and AI integrations.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to handle untrusted data, specifically when auditing code or processing external inputs. It mitigates this surface by instructing the agent on the following controls:
- Ingestion points: Processes user-provided code, external API data, webhooks, and LLM outputs (SKILL.md).
- Boundary markers: Recommends strict trust boundary mapping and schema-based validation at the system boundary.
- Capability inventory: While the skill itself is instructional, it functions in environments where agents may possess file system or shell capabilities for auditing tasks.
- Sanitization: Mandates the use of parameterized queries, output encoding, and allowlisted sanitization using tools like DOMPurify.
- [SAFE]: Explicitly warns against and provides patterns for mitigating AI-specific threats, such as prompt injection and insecure handling of model outputs.
- [SAFE]: Recommends secure development practices, including environment-based secrets management, strict CORS configurations, and dependency auditing.
Audit Metadata