luma-ray-video
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill implements significant operational safeguards. The provided Python implementation enforces an approval workflow that validates a SHA-256 request digest and checks for an explicit budget approval before making paid API calls. It uses environment variables for secret management and applies restrictive file permissions (0o600) to local state files to ensure artifact integrity.\n- [COMMAND_EXECUTION]: The skill includes a script that performs file system operations and network requests. These actions are aligned with the stated objective of managing Luma Agents API jobs and are performed using standard libraries.\n- [EXTERNAL_DOWNLOADS]: The skill interacts with the Luma Agents API endpoint (agents.lumalabs.ai). This interaction involves well-known service infrastructure and is conducted via authenticated, secure protocols.\n- [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for processing user-supplied prompts and media references for video generation. It mitigates risks by requiring explicit execution gates and providing comprehensive guidance on validating the generated output for safety and accuracy.
Audit Metadata