chengfeng-subtitle-talking-head

Pass

Audited by Gen Agent Trust Hub on Aug 1, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes several Node.js scripts (e.g., ensure-runtime.cjs, videocut-cli.cjs) to manage the subtitle workflow. It dynamically discovers the path to these scripts using the codex CLI to ensure it interacts with the correct version of the chengfeng-videocut plugin.
  • [EXTERNAL_DOWNLOADS]: The skill's runtime initialization process includes a mechanism to download and install missing components from GitHub Release. This is a standard update and installation feature for the vendor's ecosystem and targets a well-known service.
  • [PROMPT_INJECTION]: The skill ingests external data such as transcript.json, edit-list.json, and author scripts for word alignment. While this constitutes an indirect prompt injection surface, the skill utilizes specialized CLI tools for processing rather than passing raw untrusted data directly into a conversational prompt, mitigating the risk of subverting agent behavior.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 1, 2026, 10:46 AM
Security Audit — agent-trust-hub — chengfeng-subtitle-talking-head