seedance-2-5-skill

Pass

Audited by Gen Agent Trust Hub on Aug 4, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses child_process.spawnSync in scripts/generate.mjs and scripts/providers/atlas-cli.mjs to execute local utilities including FFmpeg, ffprobe, and the atlas CLI. These tools are used for intended media processing tasks such as cropping storyboards, probing media metadata, and stitching video segments.
  • [EXTERNAL_DOWNLOADS]: The documentation in references/execution-adapters.md provide instructions for installing the vendor's CLI tool via a shell script hosted on the author's official GitHub repository (https://raw.githubusercontent.com/AtlasCloudAI/cli/main/install.sh). This is a standard installation path for the author's own software.
  • [DATA_EXFILTRATION]: The skill facilitates network communication with the author's API (api.atlascloud.ai) to upload assets and manage generation tasks. It accesses local media files within the specified project directory to perform these operations, following a standard resource-handling pattern for video generation workflows.
  • [CREDENTIALS_UNSAFE]: The skill manages the ATLASCLOUD_API_KEY using environment variables. The implementation in scripts/providers/atlas-rest.mjs and the user instructions in SKILL.md emphasize secure handling, advising users to set keys in the execution environment rather than sharing them in conversation threads.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 4, 2026, 06:35 PM
Security Audit — agent-trust-hub — seedance-2-5-skill