de-slop
Pass
Audited by Gen Agent Trust Hub on Apr 13, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill functions as a text-processing utility and does not contain any malicious instructions, obfuscation, or unauthorized network operations. Access to internal resources like brand guides and voice profiles via tools (
get_marketing_reference,get_voice_profile) is consistent with the skill's purpose. - [PROMPT_INJECTION]: The skill processes untrusted user-pasted text (SKILL.md), which represents an indirect prompt injection surface. Evidence: 1. Ingestion points: user-provided text input in SKILL.md. 2. Boundary markers: absent. 3. Capability inventory: list_marketing_references, get_marketing_reference, get_voice_profile in SKILL.md. 4. Sanitization: absent. The risk is limited as the agent has no capabilities for external network communication or file-system modification associated with this text processing.
Audit Metadata