admin-portal
Pass
Audited by Gen Agent Trust Hub on May 18, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: Indirect Prompt Injection Surface. The skill documentation describes an administrative system that processes data from various untrusted or externally-controlled sources, which could contain malicious instructions.
- Ingestion points: Data enters the agent context via Client CRM leads, audit instances and responses, CI pipeline runs, and social outreach snapshots (follower_snapshots).
- Boundary markers: There are no defined delimiters or instructions to ignore embedded commands within the processed data mentioned in the skill.
- Capability inventory: The agent is instructed to build and modify features, API routes, and components that interact directly with these data sources.
- Sanitization: The documentation does not specify sanitization or validation protocols for data retrieved from the database or external API endpoints.
Audit Metadata