directus
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to fetch and process items from a CMS (Directus), which creates a potential surface for indirect prompt injection if the stored content contains malicious instructions. Ingestion points: The agent reads data using the Directus API (/items/posts) or the JavaScript SDK (readItems) as described in SKILL.md. Boundary markers: The instructions do not define clear delimiters or 'ignore instructions' warnings for the agent when processing retrieved content. Capability inventory: The skill allows the agent to create items, update items, and configure triggers for automation flows via the API. Sanitization: No explicit sanitization or filtering of external content is specified before processing retrieved data.
- [EXTERNAL_DOWNLOADS]: The skill references and uses official Directus Docker images and Node.js SDK. These are from well-known and reputable sources associated with the tool's core functionality and are treated as safe.
- [COMMAND_EXECUTION]: The skill includes instructions for deploying Directus using Docker and interacting with its API via curl, which are standard operational procedures for managing this software.
Audit Metadata