reply-to-customer-review
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted data in the form of customer reviews provided by the user. While this creates a theoretical surface for indirect prompt injection, the risk is mitigated by the following factors:
- Ingestion Points: User-provided text in the
SKILL.mdprocess flow. - Boundary Markers: The instructions explicitly separate the input analysis from the response generation, though specific character-level delimiters for the input text are not mandated in the prompt instructions.
- Capability Inventory: The skill lacks dangerous capabilities such as file system writes, network requests, or shell command execution.
- Sanitization: The skill employs a dedicated 'Humanization' step and references an external 'text-humanizer' ruleset to strip common AI writing patterns and enforce plain-text output, which serves as a functional filter for potentially malicious markdown or formatting injected through review text.
Audit Metadata