klingai-video-extension

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses subprocess.run to execute ffmpeg commands for video processing, specifically for creating seamless loops via crossfades and concatenating multiple video segments. These calls use specific, hardcoded filter chains and parameters.
  • [EXTERNAL_DOWNLOADS]: The skill fetches video content from external URLs provided by the Kling AI API using requests.get to facilitate local concatenation and processing.
  • [INDIRECT_PROMPT_INJECTION]: The skill accepts user-provided text prompts to describe video scenes and continuations which are passed to an external generation API.
  • Ingestion points: prompt and continuation_prompt arguments in SKILL.md and references/video-extension-implementation.md.
  • Boundary markers: None present in the provided scripts.
  • Capability inventory: Network requests via requests.post and shell execution via subprocess.run (ffmpeg).
  • Sanitization: No explicit sanitization of prompt strings is implemented before transmission to the external API.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 03:45 AM
Security Audit — agent-trust-hub — klingai-video-extension