talking-goose-creator-video
Pass
Audited by Gen Agent Trust Hub on Aug 19, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses
subprocess.runinreferences/composite.pyto manage video composition viaffmpegandffprobe. These operations are performed using structured argument lists to minimize command injection vulnerabilities. - [EXTERNAL_DOWNLOADS]: The script
references/generate-sfx.shfetches audio data fromapi.elevenlabs.iousingcurl. This is a standard integration with a well-known service. Additionally, this script contains a hardcoded local file path to the author's environment, which leaks metadata about the development workstation. - [PROMPT_INJECTION]: The skill accepts user-provided text for
conceptandbeats, which are interpolated into AI prompts for generating character images and voiceover scripts. This represents an indirect prompt injection surface. - Ingestion points: User-defined
conceptandbeatsinSKILL.md. - Boundary markers: None identified in the prompt construction workflow to isolate untrusted input.
- Capability inventory: Execution of shell commands and network requests.
- Sanitization: No input sanitization or filtering is applied to user-supplied text.
Audit Metadata