genmedia-workflow

Pass

Audited by Gen Agent Trust Hub on Nov 7, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill extensively uses shell commands through the genmedia CLI to interact with model registries, inspect schemas, and execute generation nodes. These commands include specific flags for JSON output and asynchronous execution to ensure structured handling of data.
  • [EXTERNAL_DOWNLOADS]: The instructions reference the genmedia-cli hosted on GitHub (github.com/fal-ai-community/genmedia-cli). As this is a vendor-owned resource corresponding to the skill author, it is considered a legitimate dependency for the skill's functionality.
  • [DATA_EXFILTRATION]: The skill uses genmedia upload to transfer local assets (images, audio, video) to the fal.ai infrastructure for processing. This data transfer is a documented and necessary part of the media generation workflow and targets the vendor's own services.
  • [INDIRECT_PROMPT_INJECTION]: The skill handles untrusted data inputs (source assets) and model outputs within its pipeline. While this presents an ingestion surface for potential indirect instructions, the risk is mitigated by the skill's focus on media transformation and the use of a structured CLI for orchestration rather than raw prompt interpolation.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 7, 2026, 06:23 PM